Professional Documents
Culture Documents
VCA-DCV
Official Cert Guide
With books, certification and study guides, video training, and learning tools produced
by world-class architects and IT experts, VMware Press helps IT professionals master a
diverse range of topics on virtualization and cloud computing. It is the official source of
reference materials for preparing for the VMware Certified Professional examinations.
VMware Press is also pleased to have localization partners that can publish its prod-
ucts into more than 42 languages, including: Chinese (Simplified), Chinese (Traditional),
French, German, Greek, Hindi, Japanese, Korean, Polish, Russian, and Spanish.
With books, certification and study guides, video training, and learning
tools produced by world-class architects and IT experts, VMware Press
helps IT professionals master a diverse range of topics on virtualization
and cloud computing, and is the official source of reference materials
for completing the VMware certification exams.
Matthew Vandenbeld
Jonathan McDonald
Special Sales
For information about buying this title in bulk quantities, or for special sales
opportunities (which may include electronic versions; custom cover designs; and
content particular to your business, training goals, marketing focus, or branding
interests), please contact our corporate sales department at corpsales@pearsoned.
com or (800) 382-3419.
For government sales inquiries, please contact governmentsales@pearsoned.com.
For questions about sales outside the U.S., please contact international@pear-
soned.com.
Introduction xviii
CHAPTER 1 Datacenter Virtualization 3
CHAPTER 2 Business Challenges 21
CHAPTER 3 vSphere Components 41
CHAPTER 4 vSphere Storage 77
CHAPTER 5 VMware vSphere Networking 101
CHAPTER 6 Virtualization Solutions That Solve Business Challenges 125
APPENDIX A Answers to the “Do I Know This Already?” Quizzes and Chapter Review
Questions 143
Glossary 153
Introduction xviii
Chapter 1 Datacenter Virtualization 3
“Do I Know This Already?” Quiz 3
Foundation Topics 6
What Is Virtualization? 6
Datacenter Components 7
Compute 8
Storage 9
Why Virtualize? 11
Optimization 11
Availability 12
Management 13
Scalability 14
Summary 15
Exam Preparation Tasks 16
Review All the Key Topics 16
Definitions of Key Terms 17
Chapter Review Questions 17
Scalability Challenges 29
What Is Scalability? 29
Common Business Challenges 31
Infrastructure Scalability 31
Virtual Machine and Application Scalability 31
Optimization Challenges 32
What Is Optimization? 32
Common Business Challenges 33
Monitoring Optimization 33
Efficiency Optimization 33
Performance Optimization 34
SMB Versus Enterprise Challenges 34
Size of the Company 34
Summary 36
Exam Preparation Tasks 37
Review All the Key Topics 37
Definitions of Key Terms 37
Review Questions 37
Availability 91
Disaster Recovery 92
vSphere Data Protection 93
vSphere Replication 94
Summary 95
Exam Preparation Tasks 96
Review All the Key Topics 96
Definitions of Key Terms 97
Chapter Review Questions 97
Answers to the “Do I Know This Already?” Quizzes and Chapter Review Questions 143
Glossary 153
To Amanda, Victoria, and Nora. Your love and support mean the world to me and have
enabled me to succeed in my aspirations.—Matt
This book is dedicated to Audrey. May she forgive my constant working and travel in
pursuit of my career and dreams. It is not often that I look back and say, “I can’t believe I
did all of that in the last year!” This is one of those times…. Sadly, I could not find a way
to incorporate a dragon into it.—Jonathan
First, thank you to Joan Murray and Ellie Bru, who tirelessly encouraged us until
we had finished the writing and editing for this book. Joan and Ellie, thank you for
not strangling us and for providing great support to a couple of new authors who
didn’t quite realize all of the work involved in getting it out the door.
We would also like to thank our technical editor, Doug Baer. It is not an easy job
translating the technical to something that can be easily consumed. Thanks for the
insight and timely reviews.
Thank you to our managers John McGinn, Rupen Sheth, and Justin McGarry, who
have been supportive of us in this endeavor to share our knowledge.
Lastly, we would like to acknowledge all of our family and friends. It takes an
amazing amount of effort to complete a book. We thank you for the tireless hours
during which we were creating and editing. This completes a journey for us moving
an item from a “wanted to do” to a “have done” on the bucket list.
Thank you!
As the reader of this book, you are our most important critic and commentator. We
value your opinion and want to know what we’re doing right, what we could do bet-
ter, what areas you’d like to see us publish in, and any other words of wisdom you’re
willing to pass our way.
We welcome your comments. You can email or write us directly to let us know
what you did or didn’t like about this book—as well as what we can do to make our
books better.
Please note that we cannot help you with technical problems related to the topic of this book.
When you write, please be sure to include this book’s title and author as well as
your name, email address, and phone number. We will carefully review your com-
ments and share them with the authors and editors who worked on the book.
Email: VMwarePress@vmware.com
Mail: VMware Press
ATTN: Reader Feedback
800 East 96th Street
Indianapolis, IN 46240 USA
Reader Services
Visit our website and register this book at www.informit.com/title/9780133854497
for convenient access to any updates, downloads, or errata that might be available
for this book.
Introduction
There is no questioning the value which virtualization has brought to the IT in-
dustry. VMware has products and technologies that revolutionize the way in which
businesses operate. From revolutionizing the way in that an administrator can
recover from failures to providing the ability to deploy a large number of servers
very quickly, the way that businesses operate is changing. The software defined
data center (SDDC) is revolutionizing the way in which IT manages the data cen-
ter. Understanding these products and technologies is instrumental to successful
implementation of them.
The value of showing your level of knowledge is second to none. Whether you
love or hate certifications, they are a necessary part of the industry. They provide
a baseline for employers to determine the level of expertise that a candidate has
in a specific area. The VMware Certified Associate—Data Center Virtualization
(VCA-DCV) certification is an entry-level certification meant to show a basic un-
derstanding of VMware data center technologies.
With the VCA-Data Center Virtualization certification, you’ll have greater cred-
ibility when discussing data center virtualization, the business challenges that
vSphere is designed to address, and how virtualizing the data center with vSphere
addresses those challenges. You’ll be able to define data center virtualization and
provide use case scenarios of how vSphere and data center virtualization can pro-
vide cost and operational benefits.
Although not a requirement to the VMware Certified Professional—Data Center
Virtualization certification, the VCA-DCV is a great starting point in your certifi-
cation journey.
Whether you simply want to become more conversant in virtualization technolo-
gies or ultimately want to be a recognized virtualization expert, VMware certifica-
tion is now an essential step for your career.
4. Click on the “Access Bonus Content” link in the Registered products section of
your account page, to be taken to the three practice exams.
This chapter covers a portion of the VCA-DCV objective 1.1 and objective 1.2.
Datacenter Virtualization
Table 1-1 “Do I Know This Already?” Foundation Topics Section-to-Question Mapping
Foundations Topics Section Questions Covered in This Section
What Is Virtualization? 1, 2
Datacenter Components 3–6
Why Virtualize? 7
5. Which two types of virtual switches are available with vSphere? (Choose two.)
a. vSphere Standard Switch
b. Consolidated Virtual Switch
c. vSphere Enhanced Switch
d. Distributed Virtual Switch
Foundation Topics
What Is Virtualization?
Virtualization is the abstraction of resources from its underlying hardware. The core
component of a server virtualization platform is the hypervisor. The hypervisor is a
piece of software that operates as a go-between for the guests’ operating systems
(OSes) and the bare metal hardware. Guest workloads are contained and isolated
from each other; this container is deemed a virtual machine (VM). The hypervisor
abstracts the physical hardware and separates it from the VM’s operating system.
This enables multiple VMs to operate on the same piece of physical hardware in
complete isolation. It also enables the guest to work on varying hardware platforms
without reconfiguration of the OS.
A hypervisor is installed on a physical server called a host. VMware’s ESXi hypervisor
is considered a type 1 hypervisor. These types of hypervisors do not require another
operating system to run on. Type 1 hypervisors run on bare metal hardware and
are very efficient. VMware hypervisors have become extremely efficient and highly
available. In most cases, there is no difference in performance for applications hosted
in vSphere versus running on bare metal. VMware’s most recent hypervisors run the
majority of workloads at near native performance while offering several other ad-
vantages that are covered later in this chapter.
Virtual machines, while operating on the same physical hardware, are completely
isolated from one another. Each VM is presented virtual hardware from the hypervi-
sor and operates in the exact same manner as if it were running on physical hard-
ware. Newer operating systems have embraced virtualization and are designed to
work efficiently as virtual machines. VMs send instructions to the virtual hardware
that are processed by the hypervisor and scheduled in a fair and efficient manner on
the physical hardware.
Figure 1-1 visualizes the core components of virtualization. The basic building
blocks of a datacenter are abstracted via the VMware vSphere hypervisor and pro-
vided to guest virtual machines. All components are then easily managed by vCenter
Server, which greatly increases the manageability of the entire datacenter.
Manage
VM VM VM VM VM VM VM VM VM VM VM VM
Datacenter Components
In this section, we discuss the core components of the virtual datacenter. These
components may differ slightly from other datacenter definitions.
The datacenter components that form the components of a virtual datacenter are as
follows:
■ Compute
■ Network
■ Storage
Server virtualization, via the hypervisor, abstracts these elements and presents virtual
equivalents to the virtual machine’s operating system. Without these core compo-
nents, a VM will not function. Although power, cooling, racking, cabling, and other
components might be referenced, they are outside the core virtual datacenter defini-
tion.
Compute
Datacenter compute components are comprised of central processing unit (CPU)
and memory resources. These resources facilitate the computational requirements
for applications. CPU and memory are housed within a physical server. This physi-
cal server becomes the host for the ESXi hypervisor. ESXi takes control of all of
the scheduling for CPU and memory resources. It efficiently and fairly shares these
resources amongst many virtual machines.
An instruction sent from a virtual machine to its virtual CPU (vCPU) is processed
by the hypervisor and scheduled on a physical CPU core. Multiple cores and mul-
tiple vCPUs can be scheduled to run concurrently, or rather, at the same time. This
concept is referred to as co-scheduling. vSphere utilizes a type of scheduling known
as relaxed co-scheduling. This type of co-scheduling flexibly schedules the vCPUs of
a guest in an intelligent manner that is an improvement in the hypervisor schedul-
ing algorithm of other hypervisors. This greatly increases the efficiency of multiple
workloads on a single piece of hardware. The hypervisor uses the concept of shares
to prioritize access to CPU and memory resources. These shares are customizable
and are used in the resource entitlement calculation. When a host is not busy and
there are CPU cycles to spare, each VM’s entitlement will equal its demand. If, for
example, all CPU cores within a host are busy because many busy VMs are trying
to use the same physical CPUs, the hypervisor uses the entitlement calculation to
determine which workload to run next or whether to interrupt a running process for
another.
Memory is allocated within the hypervisor to each VM to maintain isolation; this
allocation uses a hypervisor data structure called a pmap, which maps the “physical”
memory that a VM sees to actual physical host memory. This mapping enables the
guest OS to see a contiguous block of addressable memory, even though the actual
memory blocks may be distributed across the host’s physical memory. The ESXi hy-
pervisor uses an entity called the shadow page table to maintain virtual-to-physical
memory mapping. ESXi scans memory to find duplicate entries; these entries are
then consolidated into a single entry and pointers are created in place of the du-
plicates. This allows for oversubscription of memory resources and more efficient
use of memory as a whole. This is one of the techniques ESXi utilizes to manage
memory, called transparent page sharing. During times of contention, ESXi can use
memory reclamation mechanisms such as ballooning, swapping, and compression
Network
Datacenter networking facilitates the communication between devices both inside
and outside the datacenter. Packets are transported from the source to the destina-
tion. Dependent on where both the source and destination are located, the packet
will be moved through varying devices, both physical and virtual. The network may
consist of routers, switches, and adapters at the physical layer. Networks can also be
segmented at a logical layer into different network segments. Virtual local area net-
works (VLANs) are an example of this.
At the virtual layer the hypervisor creates virtual switches to enable VMs to commu-
nicate with the network. There are two different types of virtual switches—vSphere
standard switch (vSS) and virtual distributed switch (vDS). vDS can also be referred
to as distributed virtual switches (DVS).
Virtual network adapters are presented by the hypervisor to the guest operating
system. There are multiple types of adapters to ensure performance and support for
multiple operating system types. The guest utilizes this virtual network adapter to
send and receive packets via the hypervisor. The hypervisor controls all packets from
its physical adapters through the virtual switch to the guest, and vice versa.
Storage
Storage devices contain all data within a datacenter, including operating systems
and VMs. Data integrity and reliability are paramount within datacenters because
storage corruption or unavailability will cause major service outage. Although data
integrity is paramount, performance must not be overlooked. Storage is the most
common cause of performance-related issues within a datacenter. A balance between
data protection and performance must be achieved. The storage technology used to
achieve these goals is a Redundant Array of Independent Disks (RAID). RAID tech-
nology uses many hard disk drives to offer protection from individual disk failure
as well as offer performance via multiple devices operating concurrently. There are
multiple RAID levels, with the most common being these:
■ RAID0 Striping—Striping between drives for performance; no redundancy
for disk failures. Requires at least two drives.
■ RAID1 Mirroring—Mirroring between drives; slower performance but will
tolerate drive failures. Requires at least two drives, and drives must be added in
pairs.
Multiple other RAID types are available depending on the type of storage device
you are using. Familiarize yourself with your particular vendor’s RAID offerings to
ensure you have selected the appropriate level for your requirements.
RAID arrays can be located locally within servers or on shared storage devices. Stor-
age devices within the datacenter vary depending on the size of the organization and
the associated budget. The most common types are direct attached storage (DAS),
network attached storage (NAS), and storage area network (SAN). The most com-
mon example of direct attached storage is the disk drives located within servers.
NAS and SAN devices are shared storage devices that can be accessed by multiple
systems concurrently. NAS and SAN devices are very similar in nature—in some
cases interchangeable. The primary differentiator is how the storage is made avail-
able to hosts. SAN presents block-based storage, whereas NAS presents file-based.
These storage devices store data in one of two ways, at the file level or at the block
level. File-level storage requires the storage array to provide hosts with a file system
for them to utilize; the most common of these are NFS and CIFS. Block-level stor-
age provides hosts with raw storage—storage without a file system—the hosts them-
selves then write their own file system. ESXi formats raw devices with the VMFS
file system.
vSphere supports two types of file systems: NFS and VMFS. There are advantages
and disadvantages to both. Neither system is better than the other; the selection of
the file system is heavily dependent on your design requirements and constraints.
vSphere hosts then use these file systems as datastores. Datastores are the location
the hosts use to store VM data. vSphere datastores can be shared amongst many
hosts. This enables many of the vSphere availability features and is a key component
of the virtual datacenter.
Why Virtualize?
Virtualization has revolutionized the IT industry. In a very short time, it has gone
from small dev/test environments to enterprise tier 1 datacenters. The majority of
workloads worldwide now run on a virtualized platform, and this ratio will only in-
crease in the future. Virtualization is also the foundation for cloud platforms, which
is the future of the IT landscape. What is it about virtualization that has caused such
a rapid transformation of the IT industry? There are an incredible number of rea-
sons to virtualize; included here are a few of the most prominent reasons.
Optimization
The initial motivator to move to virtualization was server consolidation. Server
sprawl was a very big problem in the “1 app, 1 server” datacenter environment.
Hundreds of servers were using power, generating heat, taking up space, and requir-
ing administration. On top of that, the great majority of these servers were woefully
underutilized, using just a fraction of their CPU and memory. It was a completely
inefficient way to run a datacenter. Virtualization allowed one physical server to be
shared by multiple VMs. In the early days, you could get ratios of 20 to 1—this rep-
resented significant cost savings to the business. Over time, these ratios have grown,
and it is not uncommon to see ratios of 100 to 1 now.
These consolidation ratios provide a high return on investment (ROI) for purchased
physical equipment and lower the total cost of ownership (TCO) for an organiza-
tion. This is one of the primary motivators for businesses to embrace virtualization.
Virtualizing physical servers saves money that would have been spent on more phys-
ical servers than required, and gets full utilization from servers already purchased.
There are some situations in which high consolidation ratios are not preferred. This
is the case in tier 1 or mission-critical applications. These applications are being vir-
tualized to take advantage of the availability provided within a vSphere environment.
Figure 1-2 illustrates the optimization and consolidation of datacenter resources.
APP
OS
APP
APP
OS
OS APP
APP
OS
OS
APP
OS
Figure 1-2 An entire rack of servers consolidated onto a single vSphere server
Availability
vSphere enables applications to be migrated between hosts without interruption to
service, which is highly advantageous and allows for server maintenance in a greatly
simplified manner. The latest iteration of vMotion (dubbed Enhanced vMotion)
enables live migration without the requirement of shared storage. This allows you
to move workloads away from essentially any device to perform maintenance or up-
grades without impact.
In the event of unplanned outages, vSphere uses the High Availability (HA) feature
to automatically restart workloads on functioning hosts in the cluster. This greatly
reduces the impact of host failures within a datacenter. In traditional datacenters,
a host failure could interrupt application availability for hours to days depending
on the failure. However, with vSphere HA, this outage lasts a matter of minutes.
HA is also configurable, and you can select certain VMs to have a higher or lower
restart priority depending on your requirements. The newest iteration has added
VM VM VM VM VM VM
Resource Pool
Figure 1-3 Workloads moving automatically to available servers from a failed server
Management
Managing hundreds of physical servers is an onerous task further complicated by
differing and expensive management tools. VMware’s vCenter Server greatly reduces
administrative effort to manage large environments. In traditional datacenters, most
organizations tried to maintain a ratio of 1 administrator to every 25 servers. Today,
a single administrator can look after hundreds of virtualized servers with ease. This
frees up resources to drive even more projects and innovation within an organiza-
tion as opposed to just keeping the lights on.
To further simplify operations management, vCenter will automatically relocate
resources in times of contention to ensure the workload is evenly distributed. This
function is handled by the Distributed Resource Scheduler (DRS) and is highly con-
figurable and robust. The DRS concept has been extended to the storage subsystem
as well and can manage the utilization of datastores in the environment. Storage
DRS (SDRS) ensures VMs are placed onto datastores with enough capacity and
periodically checks the latency of the datastore. If specified thresholds are exceeded,
vCenter moves the VM to another datastore that will meet its requirements.
Figure 1-4 shows vMotion that greatly eases management tasks.
VM
wa
re
ES
Xi
vM
oti
on
VM
wa
re
ES
Xi
Figure 1-4 vMotion live migrating a running VM from one host to another
Scalability
Provisioning times for servers have been drastically reduced from weeks to minutes.
To provision a physical server, it must be procured, shipped, received, unboxed,
racked, cabled, an OS deployed to it, custom drivers installed, and security patched.
To provision a virtual server, an administrator can right-click a template and select
Deploy VM from This Template. It’s really that simple—as long as templates have
been created and are up-to-date. This incredible ease of provisioning and the stan-
dardization of templates have enabled the cloud principle of self-provisioning.
To scale up physical servers in the traditional datacenter, there must be an abun-
dance of room (rack space), cooling, and power. In addition, complexities exist
around procuring physical servers and their build times. With vSphere, scaling is
core to the product. Virtual machines can be rapidly provisioned and managed easily.
A greatly reduced number of physical servers are required, and ESXi can be rapidly
deployed to new hardware, especially with the use of Auto Deploy and host profiles.
These tools enable you to provision new hosts within minutes of hardware arriving
onsite; these hosts then can host many VMs.
VM
oti
on
VM
wa
re
VM
wa
re
VM
wa
re
Summary
In this chapter, you have learned the basics of what virtualization is and what its key
components are. We explained the role of the hypervisor in abstracting physical
hardware. The concept of VMs and their functionality was introduced. These are
the core concepts that build to an understanding of what virtualization is and how it
works.
The components of the datacenter were introduced in this chapter. Compute, net-
work, and storage are the building blocks utilized by virtualization. Each component
is abstracted by the hypervisor and made available to VMs to use.
The reasons virtualization are used by business were also explored. The key con-
cepts of optimization, availability, management, and scalability were explained. VM-
ware virtualization offers many benefits to businesses and has become an integral
part of the majority of datacenters.
9. Which vSphere feature enables you to migrate a VM to a new host and data-
store at the same time?
a. vMotion
b. Storage vMotion
c. Multi vMotion
d. Enhanced vMotion
10. Which vSphere feature enables the rapid provisioning of vSphere hosts?
a. Auto Deploy
b. DRS
c. Templates
d. Clones
Business Challenges
When you are operating a business, a primary concern is the common busi-
ness challenges that need to be taken into account to ensure that the day-to-
day operations are smooth and uninterrupted. In this chapter, you learn about
the types of challenges businesses face while running a datacenter. Availability,
management, scalability, and optimization of the datacenter resources are of
utmost importance to ensuring that the cost of doing business remains as low as
possible. Understanding the challenges also ensures that the recovery time and
recovery point objectives (RTO and RPOs) can be flexible and met with ease.
These challenges can vary from small businesses to large enterprises; however,
all businesses are looking to achieve the same basic needs, only on a different
scale. Specific technologies offered by VMware that enable you to meet these
business challenges are not discussed until Chapter 6, “Virtualization Solutions
That Solve Business Challenges.”
Table 2-1 “Do I Know This Already?” Foundation Topics Section-to-Question Mapping
9. Which characteristics will have a larger impact for enterprises than SMBs?
(Choose all that apply.)
a. Number of failover hosts available
b. Ensuring management network availability for hosts
c. Scale of the infrastructure
d. Number of virtual machines
Foundation Topics
Availability Challenges
To businesses, being highly available is of utmost concern. In many cases, even a lit-
tle bit of downtime can cost millions of dollars. This section discusses what availabil-
ity is, as well as the various types of challenges customers face in the environment.
What Is Availability?
Availability in its simplest form is the amount of time that a resource or service is
available to be consumed. This could be in terms of an application, an entire server,
or even an entire site. It is all relevant to ensuring that users are able to access the
resources they need to do their jobs. Figure 2-1 shows an example of a process flow
for ensuring an application is available.
2 APP X1 APP
3 OS OS
Managed Hosts
In this case, the application has a failure (1), thus a restart on the application is at-
tempted first (2), and then the operating system is completely rebooted in an at-
tempt to restore application availability (3). This demonstrates a workflow that could
be desired if there is a failure. Failures often are unique and need to be addressed in
completely different ways, depending on the cause of the issue.
Availability is also very important to the recovery time objective (RTO) and recovery
point objectives (RPOs) of the organization. Thus, an appropriate solution has to be
in place to cover many types of failures.
When looking at different types of availability, you can break it down into these cat-
egories:
■ Hardware—Where an administrator needs to take into account the potential
for failure of physical hardware and how the environment reacts to a server
that is physically having a component failure. Generally, these errors are cata-
strophic and need to be addressed before operation can continue normally.
Hardware Availability
Common hardware availability challenges often sound like the following statements:
■ “My company needs to minimize the downtime caused by a host failing in the
environment.”
■ “My company has a service level agreement that calls for a very quick response
to a host failing. Therefore, I have to be sure that virtual machines are up and
running as quickly as possible after a failure.”
■ “I need to ensure that if a host loses network access temporarily, appropriate
action is taken.”
Application Availability
Common application availability challenges often sound like the following state-
ments:
■ “We have an application that cannot be interrupted by hardware failures.”
■ “I need to ensure that if my application stops responding, action is automati-
cally taken to try to resolve the problem.”
■ “My company requires that if an operating system stops responding for a pe-
riod of time, action is automatically taken to try to resolve the problem.”
Maintenance
Common maintenance challenges often sound like the following statements:
■ “I need to be able to perform hardware maintenance at any time without im-
pacting the availability of services.”
■ “We need to avoid taking multiple servers down to apply hotfixes.”
■ “My company needs to perform hardware updates at any time of the day with-
out impacting users who are working.”
Environmental Availability
Common environmental availability challenges often sound like the following state-
ments:
■ “Our process to back up servers is difficult. It is hard to test for validity to
ensure that they are recoverable, and it is hard to restore an entire server
quickly.”
■ “My company has mission-critical applications that must be available even if
environmental conditions cause problems at the datacenter. We would like a
solution for this that will also allow us to easily test the environment to be sure
that any action that is taken will be successful if we do have a disaster.”
■ “We would like to replicate server data to another site but have different types
of storage array at each site.”
Management Challenges
Being able to manage a large environment is almost always a challenge for adminis-
trators. In this section, you learn about management and the challenges administra-
tors often face in managing an environment.
What Is Management?
Management is the ability to effectively administrate the environment and easily
work with any system in it. In addition, a single pane of glass into the environment
is a goal many companies have to address the complexities of managing an environ-
ment. Management includes such things as being able to quickly provision new sys-
tems; managing an overcrowded datacenter, roles, and responsibilities; and tracking
and monitoring for changes to the environment. This is shown in Figure 2-2.
Management Server
Manage
VM VM VM VM VM VM VM VM VM VM VM VM
Datacenter Management
Common datacenter management challenges often sound like the following state-
ments:
■ “I am the only administrator for the environment, and I need a centralized way
to manage my datacenter.”
■ “My datacenter is almost out of space. Power and cooling are maxed out, and it
is difficult to introduce new hardware into the environment.”
■ “Our environment is complex, and we have many administrators around the
world. We need to be able to see everything around the world in a single pane
of glass for administrators.”
Provisioning Management
Common provisioning management challenges often sound like the following state-
ments:
■ “We have many physical servers running on really old hardware. This hard-
ware is out of warranty, and the applications are difficult to reinstall.”
■ “I need to be able to provision a server from a standardized image that my or-
ganization uses.”
■ “My company is growing quickly, and we need a way to rapidly deploy many
virtual machines to meet the demand.”
Compliance Management
Common compliance management challenges often sound like the following state-
ments:
■ “My company is in retail, and we have a need to ensure we meet compliance
regulations for credit card information.”
■ “We need to monitor our environment to make certain that our servers are
running the latest patches. If they are not, an administrator needs to be noti-
fied.”
■ “I need to monitor an environment for unexpected changes. This will help to
determine whether change control has been bypassed for critical servers.”
Scalability Challenges
Scalability in an environment will in most cases dictate how much can be done with
a particular set of hardware. In this section, you learn about scalability and the chal-
lenges administrators face.
What Is Scalability?
Scalability is the ability of the infrastructure to handle the ever-growing load that is
generated as a result of doing business. This load could be as a result of increased
demand for resources (tax season or Black Friday are examples) or with the through-
put of an application. Scalability is often brought up in capacity discussions because
it is pertinent in discussions about load and increasing demand for resources in the
environment. Figure 2-3 shows an environment that has scaled to meet the demand
on it.
Hosts
APP
OS
APP
APP
OS
OS APP
APP
OS
OS
APP
OS
Scalability is a constraint of any environment. It does not take a long time before a
single server turns into multiple racks of servers or a single application needs to be
scaled out in an attempt to increase the scalability required for the environment. Be-
ing able to determine and assess the environment for these conditions is incredibly
important.
Scalability is also important to the RTOs and RPOs of the organization. If you are
unable to meet the objectives due to resource constraints, the consequences can be
catastrophic to the business. Therefore, the systems need to scale to meet the de-
mand of the business but still need to be able to handle an outage or an expanded
load. Many times the disaster recovery plan is last to be thought of until it is too late.
When looking at scalability, it can be broken down into these categories:
■ Infrastructure scalability—The ability of the underlying infrastructure to
handle the needs of the servers in the infrastructure. Generally, this is things
such as the amount of shared storage available, whether to use thin disk provi-
sioning, or the throughput on the network ports available in the environment.
■ Virtual machine and application scalability—The ability to handle the in-
creased demand for compute resources in the environment. Generally, this in-
volves the ability to add resources on demand, to scale out existing applications
easily, and to track where resources are being used.
Infrastructure Scalability
Common infrastructure scalability challenges often sound like the following state-
ments:
■ “We are a small company and do not have a SAN. We still want to be able to
take advantage of virtualization.”
■ “My company has the need to scale out the network bandwidth available to my
virtual hosts without taking downtime on the virtual machines.”
■ “I need to ensure that I can control the characteristics of the storage that is
used to ensure that I have enough room to scale my environment appropri-
ately to meet the needs of my servers.”
Optimization Challenges
Ensuring that the environment is able to effectively operate given a set of parame-
ters is key to ensuring that resources are not being ineffectively used. In this section,
you learn about environmental optimization and the challenges administrators face
in regards to it.
What Is Optimization?
Optimization is the ability to recognize inefficiencies in an environment and take ac-
tion to correct them. This enables a workload to take full advantage of the available
performance of the environment. Optimization can take place either automatically
or manually. Figure 2-4 shows a virtual machine (VM) migration taking place to bet-
ter optimize the environment by balancing the number of VMs across all hosts in
the cluster.
VM VM VM VM VM VM VM VM VM
In Figure 2-4, the system detects that one host is underused as compared to the
other hosts in the cluster. A migration (1) is done to distribute the load more effec-
tively to optimize the performance of the overall system.
Automated detection and remediation of the environment enable the environment
to be proactively monitored and action taken where necessary. This could be as sim-
ple as alerting the administrator or more complex, such as performing a migration.
Resource utilization is a common area where administrators often can benefit from
proactively monitoring the environment in this way.
When reviewing optimization challenges, they can be broken down into these cat-
egories:
■ Monitoring—Active review of the environment to determine the state of the
environment. This includes areas such as health, efficiency, and risk. These are
key areas that are important to monitor for negative future impact.
■ Efficiency—The ability to more effectively use resources to meet the growing
demand of the business without negatively impacting other servers.
■ Performance—Ensuring that the environment is optimized for peak perfor-
mance as needed by the environment. This includes prioritization of specific
resources as required by the business.
Monitoring Optimization
Common optimization monitoring challenges often sound like the following state-
ments:
■ “We would like to be able to proactively monitor for performance bottlenecks
in our environment. Currently our environmental monitoring provides only a
few basic statistics that we can look at.”
■ “My company needs to have detailed performance reports that can be used to
trend our environment.”
■ “I have several servers that are idle most of the time. I need a way to monitor
them and assess whether consolidating them will affect performance.”
Efficiency Optimization
Common efficiency optimization challenges often sound like the following state-
ments:
■ “We have several servers that have very little disk usage but were configured
with a large amount of disk space. It would be great to be able to minimize the
impact this has on the overall disk usage.”
■ “I need a way to detect whether I am efficiently using the available resources
in my environment.”
■ “My company has several different storage arrays that are used in the environ-
ment. Some of these storage arrays are high performance and cost a lot more
to operate than others. Many times development users place virtual machines
on the incorrect storage device, causing degradation of performance for the ar-
ray, which we correct by moving the virtual machine to the right class of stor-
age. We need an automatic way to ensure that users can choose the appropriate
class of storage from the time of deployment.”
Performance Optimization
Common performance optimization challenges often sound like the following state-
ments:
■ “I need a way to proactively ensure that I have a healthy environment so that I
can get optimal performance of my hosts.”
■ “We need a way to prioritize network traffic to ensure that our mission-critical
virtual machines get the peak performance possible.”
■ “My company needs to ensure that we can add resources to virtual machines in
our environment on-the-fly to address performance problems if they occur.”
Most solutions can, with some predictability, be used in any environment; others
are not suited for certain environments or for certain levels of service. For example,
vSphere Replication is a suitable technology to implement only if the RPO of the
business is less than 15 minutes. Knowing when a solution fits a particular business
size or use case is incredibly important to ensuring a properly operating environ-
ment.
In addition, the total amount of data that must be managed often greatly differs
given the size of the environment. If the rate of change of data is greater than the
possible bandwidth available to replicate that data, then it is not going to be able
to meet the needs of the company. Thus, another solution will be needed. Whether
that solution is to add more bandwidth or to be more selective as to which data is
replicated, there must be a compromise. This compromise is typically in the form of
cost versus function because the price for some solutions might not be viable for all
companies.
The following is a list of common challenges that might need to be taken into ac-
count when assessing the environment:
■ Amount of available bandwidth—In smaller businesses, the bandwidth con-
cerns might not be as much of an issue as in a larger enterprise where there is
a heavier workload. The cost of additional bandwidth can be more than is pos-
sible for the function being performed.
■ Backup strategy—It is easy to build a solution that says “Back up the entire
environment every night”; however, this might not be possible in an enterprise
that has several thousand VMs. Often the amount of time becomes prohibitive
to this type of solution.
■ Availability strategy—In an enterprise, it is a lot easier to ensure that there
are available resources to recover from a failure. Whether the failure is a single
host or an entire site, resources are available. In a SMB, however, losing a
single server could be reducing capacity by a significant amount. Thus, appro-
priate solutions and policies must be set to ensure proper recovery.
■ Centralized management—It is easy to have everything centrally managed in
a smaller business because fewer hosts typically must be managed. When you
are an enterprise business with multiple sites and many more servers, this is
a lot more difficult to accomplish. In multisite scenarios, network bandwidth,
latency, and redundancy become larger concerns than in smaller, single-site
environments.
■ Compliance and configuration management—In an SMB, keeping tabs on
environment changes that would violate compliance regulations can be easier,
and specialized software might not be required. In enterprise environments,
These are only a few examples to make you start to think about the types of differ-
ences between SMBs and enterprise environments. Some might amount to nothing,
while others might end up being a deal-breaker for the technology.
Summary
In this chapter, we discussed the various types of challenges an administrator must
consider when designing and implementing an environment. Availability, manage-
ment, scale, and optimization of the environment are key to being able to design an
infrastructure that meets the standards for the business:
■ Availability enables an administrator to take into account any environmental
failures, from a single server up to the entire site. Ensuring that you are able to
meet the RPO or RTO appropriately for the workload will dictate how avail-
able the environment needs to be.
■ Management is specific to managing the resources in the environment. Cen-
tralizing the management of resources and the ability to perform maintenance
without impact to resources are often the cardinal needs of the business.
■ Scalability of the environment is all about what is needed to be able to do the
work in the environment. In many cases, this equates to how much work can
be done in the datacenter before resources are taxed to the maximum possible.
■ Finally, optimization is the ability to monitor the environment for the health
of the environment and any anomalies that might occur during regular opera-
tions. The datacenter will have peaks and valleys in terms of utilization, to
which an administrator needs to ensure that the health does not decline.
How these challenges can be met is discussed in Chapter 6, after you learn how vir-
tualization concepts can be applied to meet these challenges
Review Questions
The review questions section is for you to review your knowledge of the topics in
the chapter. You can find the answers in Appendix A.
1. Which of the following is an availability challenge?
a. Centralized management of a few hosts.
b. Ensuring that downtime is minimized.
c. We want to use our disk space as efficiently as possible.
d. Trending on resource utilization.
2. Your manager has asked you to ensure that the mission-critical virtual ma-
chines in the environment are protected in case of environmental disaster.
Which type of challenge is this?
a. Scalability
b. Management
c. Optimization
d. Availability
4. Your boss wants to reduce the power and cooling requirements of the datacen-
ter. Which type of challenge is this? (Choose all that apply.)
a. Management
b. Scalability
c. Optimization
d. Availability
5. You need to be able to add more RAM to a VM. Which type of challenge is
this?
a. Optimization
b. Availability
c. Scalability
d. Management
8. You need to ensure that the datacenter is healthy. Which type of challenge is
this?
a. Scalability
b. Optimization
c. Management
d. Availability
vSphere Components
Table 3-1 “Do I Know This Already?” Foundation Topics Section-to-Question Mapping
2. The operating system in a virtual machine requires that specialized drivers are
installed, after the server has been successfully booted.
a. True
b. False
4. Once configured, vCenter provides which functions? (Choose all that apply.)
a. The ability to create virtual machines
b. High availability for the virtual machines
c. Dynamic resource scheduling of the virtual machines
d. Centralized management of resources
6. Resource pools enable the configuration of which resources? (Choose all that
apply.)
a. Memory
b. High availability
c. Power consumption
d. Processor
10. VMware Data Protection (VDP) can only replicate to other VDP nodes.
a. True
b. False
12. vSphere Replication is an ideal solution for an environment that has an RPO
of less than 15 minutes.
a. True
b. False
14. P2V migrations are the process of converting physical servers to virtual ma-
chines.
a. True
b. False
15. The VMware vCenter Operations Management Suite provides metric roll-up
badges for each element in the environment. What are the three main badges?
a. Health
b. Efficiency
c. Risk
d. Availability
Foundation Topics
Virtualization Concepts
It is important to understand the underlying concepts of virtualization before you
can understand how it can benefit the environment. In this section we discuss the
basic principles including virtual resources, virtual machines, and hypervisors. This
will build a framework to learn about the other technologies that increase the value
proposition of virtualization and also begin to show why it has become so popular.
Virtual Resources
In Chapter 1, “Datacenter Virtualization,” virtualization was introduced to provide
a background for why it is so popular in the business and technical worlds. VMware
uses a hypervisor to take the resources available in a physical server and abstract
them. These abstracted resources are known as virtual resources and are the core
building blocks of virtual machines. Figure 3-1 shows the four primary abstracted
resources.
NOTE CPU and RAM are often combined and called compute resources in discus-
sions that are not specific to a particular resource.
Processor
The processor in a virtualized environment is used in the same way as in any com-
puter. It is the hardware that carries out the main processing for the VM. A virtual
CPU is a required component of a VM. The difference primarily lies in the fact that
the physical central processing unit (CPU) in the host server is managed by the hy-
pervisor, so that it can be shared amongst all the VMs running on the hypervisor.
VMs are also able to utilize more than a single CPU. As of vSphere 5.5, with the
use of virtual symmetric multiprocessing (SMP), each VM can use a maximum of 64
CPUs. The hypervisor therefore must be efficient at scheduling the access to CPU
resources to ensure peak performance of all running VMs.
In addition, ESXi hosts offer the ability to hot-add vCPUs to running VMs (as long
as the guest operating system supports it) to enable expansion of resources without
experiencing downtime.
Memory
The random access memory (RAM) in a virtualized environment is also used in the
same way as a physical computer. RAM is used to store the data that is currently
being used by the running operating system and its applications. It is volatile, so if
power is lost, what was in the memory is also lost. Virtual RAM is a required compo-
nent of a VM. The primary difference in the virtual environment is that the physical
RAM in the host server is abstracted so that it can be used by all the VMs running
on the hypervisor. As of vSphere 5.5, each VM can use up to 1TB of memory.
Due to how RAM is virtualized, the hypervisor is able to allocate more RAM to
VMs than exists in the physical host. This is done by creating a swap file on disk that
is used as RAM in the case that not enough physical RAM is available. This is called
over-commitment of resources. If resources are severely over-committed, swapping to
disk will occur more frequently and can have a significant impact on performance.
In addition, ESXi hosts offer the ability to hot-add RAM to running VMs (as long
as the guest operating system supports it) to enable expansion of resources without
experiencing downtime.
Hard Disk
Virtual machine hard disks (VMDKs) are used as permanent storage for VMware
VMs. In a virtualized environment, the “hard disk” is a set of files that exist on a
physical storage device—rather than an actual physical device. The device storing
the VMDKs could be either local or remote to the host. Commonly, VMDKs are
stored on a storage area network (SAN) to enable accessibility from multiple hosts.
The benefit to this is that it opens up a completely new way to manage servers due
to the fact that the disk is just a file on a datastore. With technologies such as thin
provisioning, snapshots, and linked clones, disks might contain only a small amount
of changed data, thus reducing the storage footprint for VMs and providing new
ways to provide data recovery in your environment.
Hard disks are not a hard requirement for a VM, although most VMs have them. It
is possible to boot a VM with no hard disk, in which case it must boot from another
device. As an example, this could be a boot from the network in which all files are
loaded into to memory in a stateless configuration.
In addition, you can also have multiple disks in a VM. This provides the ability to
separate operating system and data disks, as is common in many enterprise environ-
ments.
Network
In the virtual environment, networking is largely a software construct. The network
cards in a physical host are presented to the hypervisor and assigned as uplink ports
to a virtual switch (vSwitch). You then create portgroups on the switch to provide
networks that are utilized for functions in the environment. These portgroups can
be assigned for specific types of traffic, such as host management or VM.
This construct enables increased control over data security and performance over
the virtual networks. Functions such as Quality of Service (QoS) and network I/O
control are used here to prioritize traffic to ensure peak performance.
Network cards are a requirement for management of the hypervisor; however, VMs
do not have to have them. You can have multiple network cards in both hosts and
VMs.
Virtual Machines
Virtual machines are software-based containers that enable abstraction of a physical
server’s resources. As shown in Figure 3-2, without virtualization, the operating sys-
tem and applications consume the entire physical host, regardless of actual resource
utilization. Many times the application does not fully utilize the resources of this
host. This is not only a waste of resources, but also a waste of datacenter rack space
and—more importantly—power.
APP APP
OS
OS
1 3
By using VMs, you are able to abstract the physical resources and use them more
efficiently. This is because the hypervisor has control of the underlying resources.
Therefore, rather than having an idle server, many VMs can run and use the same
physical hardware simultaneously.
VMs are a core building block of the Software Defined Datacenter (SDDC) that en-
ables you to take existing physical servers and repurpose and consolidate them. VMs
are exactly the same thing as a physical server but use virtual hardware rather than
the underlying physical server hardware. Due to this, they enable all the benefits
provided by virtualization, such as high availability or dynamic resource scheduling.
This is shown in Figure 3-3.
APP
OS
APP
APP
OS
OS APP
APP
OS
OS
APP
OS
Hypervisors
The hypervisor is what makes virtualization possible. The hypervisor is the founda-
tional software that creates and runs VMs. There are two basic types of hypervisors:
■ Bare-metal hypervisors—Run directly on the physical hardware. These in-
clude VMware ESXi and Microsoft Hyper-V.
■ Hosted hypervisors—Run on top of another operating system.
The downside of hosted hypervisors is the added level of a host operating system
that significantly reduces the performance of the VMs because the hypervisor
doesn’t have ultimate control of the resources. These include VMware Workstation/
Fusion and Parallels Desktop.
Figure 3-4 shows the various types of hypervisors.
VM VM VM VM VM
VM VM VM VM VM Hypervisor
ESXi Hosts
VMware ESXi is the current implementation of VMware’s bare-metal hypervisor. It
was first released in 2001 and significantly enhanced the performance and reliability
of x86 virtualization. At the time, there was only hosted virtualization. In the years
since, it has seen several changes all that moved toward improving the current ESXi
architecture. In the current architecture, ESXi is a fraction of the size of previous
versions. More importantly, it dramatically improves management, security, deploy-
ment/configuration, and administration of the hypervisor.
ESXi hosts are a core component to the vSphere suite and enable virtualization to
be possible at the most fundamental level. They are the component on which the
VMs run and also perform many of the infrastructure management operations.
ESXi Hosts, however, are limited until it is a managed resource in VMware vCenter
Server. Once managed, a single ESXi server is no longer just another server and can
take advantage of pooled resources, high availability, and many other functions that
are fundamental to the vSphere suite. These functions and more are discussed later
in this book.
Others
Although VMware was the pioneer to bare-metal x86 virtualization, others are avail-
able in the market as well. A few of them are listed here for reference:
■ Microsoft Hyper-V—Released in 2008, Microsoft entered into the bare-
metal hypervisor market. Being integrated with Microsoft Windows, it allows
administrators who are familiar with Windows administration to easily take
advantage of virtualization in their environments.
■ Citrix XenServer—Xen originated as a research project at the University of
Cambridge, and a company, XenSource, was created to support the project.
The first public release of Xen occurred in 2003. XenSource was acquired by
Citrix in 2007.
■ KVM—Released in February 2007, kernel-based VM (KVM) is an open-
source hypervisor built in to the Linux kernel.
Managing Resources
Now that you understand the fundamentals of how virtualization works, we can
introduce the concepts of managing the virtual resources. Being able to manage
the virtual resources can be more complex than a physical environment with the
technologies in this section. VMware vCenter Server, clusters, resource pools, and
Distributed Resource Management all enable an administrator to manage the envi-
ronment. Many of the concepts are also unique to virtualized resources, effectively
shifting when and how specific functions are performed.
vCenter Server
VMware vCenter Server is the component of the vSphere suite that ties all of the
other components together. It provides a centralized platform for managing vSphere
environments, automating many complex tasks. After a host is added to the vCenter
Server inventory, vCenter provides the following functions:
■ Simple deployment—Templates, host profiles, cloning VMs, and vApps.
■ Centralized control and visibility—vSphere Web Client to manage the en-
vironment from any browser, vCenter Single Sign-On for authentication, cus-
tomized roles and permissions, and inventory searching.
■ Proactive optimization and availability—Resource management, Distrib-
uted Resource Scheduler (DRS), Distributed Power Management (DPM),
VMware high availability (HA), and vSphere App HA.
■ Management—vCenter Orchestrator to automate management tasks, perfor-
mance data, task and event logs, and API access.
Figure 3-5 shows vCenter and how it is central to managing a VMware environ-
ment.
VMware
vCenter Server
Automation Visibility
Unlocks the power Deep visibility into
of VMware vSphere every level of the
through proactive virtual infrastructure
management
Scalability
Scalable and Extensible
management platform
VMw
are v
Sphe
re
Clusters
When adding a host to vCenter server, you must add it somewhere in the hierarchy.
The hierarchy in most cases will have at least one datacenter that contains either in-
dividually managed hosts or clusters of hosts.
Managed Hosts
A managed host is a host that is managed by vCenter but is not a part of a cluster.
Managed hosts can take advantage of central management (Single Sign-On access
control, logs, and basic host monitoring) and migration technologies (vMotion,
Storage vMotion, and Datastore Clusters). They cannot, however, participate in
functions such as VMware DRS, high availability, or virtual SAN (VSAN) activities.
Clusters of Hosts
A cluster of hosts enables vCenter to further abstract the resources of member ESXi
hosts into a pool of CPU, memory, and storage resources. As of vSphere 5.5, a single
cluster can contain up to 32 hosts. Once created, all the features of a cluster are
available to configure. These features include the following:
■ DRS
■ DPM
■ HA
■ Enhanced vMotion Compatibility (EVC)
■ VSAN
Resource Pools
Resource pools are an abstraction of CPU and memory resources. This abstraction
is used to provide a hierarchy and partitioning of the available resources into guar-
antees for VMs that are under the resource pool.
Resource pools can be configured on only standalone/managed hosts and DRS-
enabled clusters.
VM VM VM VM VM VMotion VM VM VM VM
Resource Pool
Physical Servers
While enabled, DRS monitors CPU and memory resources and generates recom-
mendations as to whether a migration to a different host will benefit the perfor-
mance. These recommendations are assigned a priority rating that can be compared
against a threshold set by the administrator for how conservative or aggressive DRS
is in automatically acting on the recommendations. DRS can be configured in three
ways:
■ Fully Automated—DRS generates recommendations for VM migrations and
initial placement and acts upon them in an automated manner; no administra-
tor intervention required.
■ Partially Automated—DRS automatically chooses where to initially power on
a VM but does not automatically act on recommendations for migrations. Ad-
ministrator intervention is required to apply the migration recommendations.
■ Manual—The administrator needs to choose where to power on a VM and
also apply any recommendations generated by DRS.
In any case, DRS is a powerful tool to ensure that there is automated monitoring for
workload contention in the environment, even if it is only acted on manually. It will
ensure that if resources are unbalanced, action will be taken.
the workloads and still satisfy availability constraints. After the workloads have been
consolidated, DPM shut downs hosts into standby mode. This is shown in Figure
3-8.
VMotion VM VM VM VM VM VM VM VM
Resource Pool
for complex failover solutions that are tied to specific applications or operating sys-
tems. VMware HA as shown in Figure 3-9 provides the following benefits:
■ Helps to minimize unplanned downtime and service disruption.
■ Eliminates the need for dedicated standby hardware.
■ No additional software installation is needed on VMs.
■ Uniform availability across the virtual environment.
VM VM VM VM VM VM
Resource Pool
After it is enabled, VMware HA continuously monitors the ESXi hosts in the cluster
for failures. This enables the environment to be proactively monitored for failures.
If a failure is detected, affected VMs are restarted on another one of the remaining
ESXi hosts in the cluster. VMware HA can detect full ESXi host hardware failures,
network isolations, and guest operating system failures.
VMware HA detects failures in the following ways:
■ Network isolation—If network connectivity fails on the management
network(s) that the agent uses to communicate between ESXi hosts, a host is
considered isolated.
■ Datastore heartbeats—If the network connectivity is detected as isolated,
the agent then checks to see whether the ESXi hosts are still accessible via the
datastores being used for heartbeats. This enables HA to determine if the host
is truly isolated or if the host has failed completely.
Each cluster has the ability to protect up to 32 hosts. You can, however, have mul-
tiple clusters protected. This enables the environment to be proactively monitored
for failures to ensure the environment is available.
No Reboot
Seamless Cutover
Hyperic App HA
Virtual Virtual
Appliance Appliance
VM VM VM VM
Agent Agent Agent Agent
The agent subsequently monitors the VMs reporting back to the Hyperic server,
which can act on changes configured in the configured App HA policy. App HA pol-
icies are configured by the administrator to match business requirements for the ap-
plications. App HA can therefore improve application uptime through the following:
VDP
Virtual
VM VM VM Appliance
VMware vSphere
VM VM VM VM VM VM VM VM VM VM VM VM
Servers Servers
Storage Replication
These benefits ensure fast and predictable recovery point and time objectives (RPOs
and RTOs) for an environment and can use existing underlying infrastructure if re-
quired.
vSphere Replication
vSphere Replication (VR) is a feature of the vSphere platform that allows an admin-
istrator to continually replicate a running VM to another location for disaster recov-
ery purposes. This enables rapid restoration of VMs in case a recovery is required. It
also enables flexible RPOs from 15 minutes all the way to 24 hours.
IMPORTANT If the recovery point objective is less than 15 minutes, vSphere Rep-
lication is not an appropriate solution.
After the images are taken, vSphere Replication can be used for restoring the images
through the vSphere web client or with a full disaster recovery product such as Site
Recovery manager. This is shown in Figure 3-14.
VM VM VM VM VM
VM VM VM VM
Migration
Virtualization also brings in technologies that enable movement of the virtualized
resources in the environment. VMware vMotion, VMware Storage vMotion, and
Physical to Virtual migrations are introduced in this section. These technologies en-
able unique new workflows in the environment, allowing maintenance windows in
the middle of the day to physical hosts and balancing of resources across the physical
infrastructure.
vMotion
VMware vMotion is a feature of vSphere that enables you to perform Live Migra-
tions of VMs between two separate physical ESXi hosts. vMotion is performed at
the hypervisor layer; therefore, it can be performed on almost any VM, regardless of
the guest operating system.
This process enables an ESXi host server to be cleared of running resources without
incurring any downtime at all for the running VM workloads. During the process,
the VM seamlessly transfers its running execution state and active memory over a
dedicated network link.
NOTE During a vMotion, the VM operating system does not realize that any-
thing is happening to it in the background. Thus, network connections proceed
normally and, other than a brief quiesce where generally no more than a single
ping is lost, the VM remains available to end users.
VM
wa
re
ES
Xi
vM
oti
on
VM
wa
re
ES
Xi
vMotion relies on having identical (or very close to identical, most notably for the
CPU) underlying physical hardware on which the ESXi host is running. Thus, live
migrations are checked for host compatibility and allowed only between hosts that
can run the VM in the current execution state.
As a result, live migrations are not allowed between Intel and AMD CPUs (the ar-
chitecture is completely different) and are disallowed if networking and/or storage
elements are not shared between hosts.
Storage vMotion
VMware Storage vMotion expands on traditional vMotion technology by allowing a
VM to be live migrated not only to a different host, but also to a different datastore.
It enables an administrator to perform a migration of disks between datastores re-
gardless of the storage array or the class of storage, as shown in Figure 3-16.
Using storage vMotion, you can ensure that VMs are on the proper class of storage
for peak performance, as well as to utilize the underlying storage more efficiently.
In addition, much like with vMotion technology, VMs can be migrated off specific
datastores or storage arrays, so that maintenance can be performed on the under-
lying storage itself without needing to take down time on the VMs.
VM VM VM VM
VMware ESXi
Storage vMotion
NOTE Much like with vMotion, the VMs themselves are unaware that a Storage
vMotion is occurring. As a result, there is no downtime to the VM when one oc-
curs.
Physical to Virtual
Physical-to-virtual migrations can also be performed with VMware vCenter
Converter. During this process, a physical Linux or Windows server, as well as
third-party image formats, can be converted into VMs that can run on VMware hy-
pervisors.
VMware Converter enables you to quickly migrate existing environments to virtual-
ized ones. This enables all applications in the environment to take advantage of the
benefits of virtualization, including the live migration and HA solutions offered by
VMware vCenter Server.
vCenter Operations Manager uses these rolled-up metrics to enable a large number
of metrics to be presented in score form, dramatically simplifying traditional moni-
toring. By adding vCenter Operations Manager to the environment, the administra-
tor will add the following functions:
■ Automated roll-ups of key metrics into simple score metrics. This benefits the
administrator because there is no requirement to know what each and every
metric does to get an overview of the environment.
■ Calculation of a normal behavior threshold for every metric. This allows the
administrator to highlight abnormalities in the environment and take proactive
corrective actions.
■ Graphical representation of complex environmental stats.
■ Display of information about changes to the environment, such as adding
workload to an ESXi host and the resultant change to the health, risk, and ef-
ficiency of the host.
Summary
In this chapter, we discussed the vSphere components that not only make virtualiza-
tion a reality, but also transform the way in which businesses operate. With virtual-
ized resources, we are able to power VMs through the use of the hypervisor and
ESXi, in the case of VMware software.
After you have virtualized the environment, it can be managed by VMware vCenter
Server and its surrounding technologies. This gives an administrator a single-pane-
of-glass view of the environment and enables allocation of the resources dynamically
in the environment using clusters, resource pools, and DRS.
However, any environment needs to be properly protected from failure, and in the
availability section, you learned about HA, DRS, and App HA. These enable indi-
vidual resources to be protected. In addition to this, you learned about VDP, SRM,
and VR—technologies that enable a larger failure to be overcome or recovered from
as quickly as possible.
vMotion technology enables the administrator to take this a step further by provid-
ing the ability to live migrate VMs. This enables the administrator to proactively
perform maintenance or decommission a physical server, as required, without im-
pacting the availability of the virtual environment.
Finally, you learned about other technology that helps to monitor the environment.
VMware vCenter Operations Manager enables the administrator to monitor the
health, risk, and efficiency of the environment. On the other hand, vCenter Config-
uration Manager enables the administrator to monitor for compliance to a standard.
In any case, the technologies offered by VMware enable new ways to manage the
environment as well as let an administrator revolutionize how change control is
done in the environment. All of this is made possible by the basic constructs of virtu-
alization.
Review Questions
The review questions section is for you to review your knowledge of the topics in
the chapter. You can find the answers in Appendix A.
1. What is the benefit of virtualizing resources? (Choose two.)
a. Consolidation of resources.
b. Reduction of datacenter power footprint.
c. More physical hardware is required.
d. Business processes stay the same.
2. How does being virtualized affect the operating system in a VM? (Choose
two.)
a. The operating system is aware that it is virtualized.
b. The VM can be scheduled to use any CPU in the host.
c. The operating system is unaware that it is virtualized.
d. The VM must be tied to a physical CPU(s).
3. What are the benefits of using a bare-metal hypervisors? (Choose all that
apply.)
a. They run on top of another operating system.
b. They control the physical compute resources in the server.
c. They run as their own operating system.
d. The administrator configures specific amounts of compute resources for
the hypervisor.
4. Managing the environment with vCenter enables the usage of which addi-
tional vSphere features? (Choose all that apply.)
a. Templates
b. VMware HA
c. VMware DRS
d. VMs
5. Clusters are required for which features? (Choose all that apply.)
a. VMware HA
b. Resource pools
c. VMware DRS
d. vCenter Operations Manager
6. Resource pools enable the configuration of which resources? (Choose all that
apply.)
a. CPU affinity
b. High availability
c. Shares
d. Reservations
8. Which term best describes the type of availability that fault tolerance
provides?
a. Continuous availability
b. High availability
c. Replicated availability
d. Distributed availability
11. VMware Site Recovery Manager is ideal for protecting against which type of
situations?
a. Power failure of a single host
b. A flood at the datacenter
c. Air conditioning failure in the datacenter
d. Installation of new ESXi servers to an existing environment
12. What is the minimum recovery point objective that vSphere Replication is
supported to use?
a. 5 minutes
b. 15 minutes
c. 30 minutes
d. 1 hour
14. Which roll-up badge in the vCenter operations management suite provides a
view of how well the resources are being used in the environment?
a. Health
b. Efficiency
c. Risk
d. Availability
vSphere Storage
Table 4-1 “Do I Know This Already?” Foundation Topics Section-to-Question Mapping
1. Which type of disk does not use spinning platters, increasing performance
drastically?
a. SCSI
b. SSD
c. SAS
d. SATA
5. What is a datastore?
a. A shared storage device
b. A collection of disks aggregated together for virtual machine storage
c. A virtual hard disk drive
d. A logical container used to store virtual disks
8. Which vSphere feature ensures datastore latency does not surpass a defined
threshold?
a. NIOC
b. SDRS
c. SIOC
d. DRS
10. Which VMware product is commonly used by SMB and branch offices to ag-
gregate local disk drives into a redundant shared datastore?
a. VSA
b. VVOLs
c. VAAI
d. VRFC
11. Which vSphere feature enables you to move a virtual machine’s data from one
datastore to another without interruption?
a. vMotion
b. Storage vMotion
c. Super vMotion
d. Storage replication
12. Which vSphere product can back up and restore entire virtual machines?
a. vSphere Data Watch
b. vCenter Backup Utility
c. vSphere Data Protection
d. vSphere Data Guard
Foundation Topics
SAN Switches
In this section, we discuss the types of storage used by the vSphere platform.
The types of storage are the following:
■ Direct attached storage
■ Network attached storage
■ Storage area network
Each storage type has different features and use cases, as well as different cost and
skills required to administer it.
enabled both a host vMotion and storage vMotion to occur at the same time, which
became known as Enhanced vMotion. This allows multiple hosts with nonshared
storage to operate in a similar manner to shared storage. The only real caveat to that
is that if a host fails, there is no way to move the VMs on that host to another server
until the server itself is available again.
Storage Protocols
Numerous storage protocols are supported by vSphere, so the selection of protocol
is highly dependent on the environment and its design. The supported protocols are
as follows:
■ Fiber Channel
■ iSCSI
■ Fibre Channel over Ethernet (FCoE)
■ NFS
■ Local
Each protocol has strengths and weaknesses; no single protocol is better than an-
other. Choose what is appropriate for your environment and your business require-
ments.
Filesystem Types
A filesystem is the manner in which a host creates a layout on which to store data.
Two primary types of filesystem are available to vSphere: VMFS and NFS. VMware
created VMFS for the sole use of vSphere, beginning with ESX, to store VMs. NFS
is a distributed filesystem protocol in use by many types of operating systems and
hypervisors. There are advantages and design considerations to each.
NFS
NFS is a distributed filesystem protocol created and maintained on the network at-
tached storage device itself. The hypervisor does not control the filesystem of the
storage, nor does it need to. The storage device manages the filesystem and the
placement of data on it.
NFS provides many advantages in virtual datacenter design, such as the ability to
mount the filesystem on multiple operating system (OS) types concurrently. This
can aide in certain tasks such as backup and recovery. NFS has historically provided
larger datastores and more data visibility compared to older versions of VMFS.
VMFS has significantly closed that gap as of VMFS-5.
VMFS
Virtual Machine Filesystem (VMFS) was created by VMware to efficiently store and
access VMs. VMFS is written on block-based directly attached or shared storage de-
vices. VMFS is VMware’s cluster filesystem that allows multiple hosts to access the
same filesystem at the same time. It enables vSphere hosts to write and read from
the datastore simultaneously by using an advanced file-locking mechanism.
VMFS has undergone several version revisions, each with new functionality. Cur-
rently, VMFS is at version 5; today’s vSphere platform supports VMFS-3 and
VMFS-5. VMFS-3 should be used for compatibility only with ESXi hosts under ver-
sion 5.0. VMFS-5 allows for a unified 1MB block size as well as a 64TB maximum
datastore size.
vSphere Datastores
vSphere datastores are logical units of storage available for VMs, VM templates,
ISOs, and floppy images. Datastores utilize either the VMFS or NFS filesystem
dependent on the background storage. These datastores can be shared with multiple
hosts to a limit of 64 for VMFS and 256 for NFS (ESX5+), or they can be created
locally on a single host. The accepted best practice is to share datastores to all hosts
in a compute cluster. This enables VMs to move between all hosts easily and in-
creases availability in the event of a host failure. Figure 4-2 illustrates a datastore in
a vSphere environment.
VMDK
A VMDK is essentially the hard disk drive of a VM. All data written and read for a
VM resides in the VMDK file. This VMDK file is stored on a vSphere datastore and
can be accessed by multiple hosts. A VMDK can be stored on any supported type of
storage device or filesystem. It is the standard building block of virtual storage.
A VMDK file can be provisioned in three ways:
■ Thin
■ Lazy-zeroed thick
■ Eager-zeroed thick
Thin-provisioned VMDKs allocate only the amount of space required for actual
data on a datastore. That means if you have a 20 GB C: drive but are using only
10 GB, it will consume only 10 GB of storage on the datastore. This type of
provisioning uses the least amount of storage and allows you to over-allocate datas-
tores. It does introduce some risk because if rapid data changes occur and you fill
up a datastore, the VMs on that datastore will cease to function. When new data is
written, the disk must first be grown, then a zero written, then the actual data. This
incurs the highest performance penalty of all the options but consumes the least
amount of space.
Lazy-zeroed thick provisioning reserves the entire 20 GB of space you have speci-
fied, but it does not write any data to the “empty” space. When new data is written,
a zero must be written first, then the actual data bit. This does incur a very small
performance penalty.
Eager-zeroed thick provisioning reserves the entire space allocated and proactively
writes zeroes for every block in the allocation. Eager-zeroed provisioning is recom-
mended for high I/O workloads such as SQL servers but consumes the most disk
space.
RDM
An RDM is a mechanism that enables vSphere VMs to access raw physical storage.
vSphere uses a VMFS volume with a mapping file to proxy commands to the raw
storage device; this allows for some vSphere features to be available (vMotion, HA)
while still offering direct access.
While not usually recommended, in certain use cases RDMs become advantageous.
Examples of this are when a cluster is created between a VM and a physical machine,
or when a specialized snapshot utility is required for backups.
SDRS
Storage DRS is a feature that can be enabled on datastore clusters and allows you
to select to have SDRS automatically perform storage actions for you or alert you
so that you can do them manually. Regardless of whether you intend to use SDRS
in automatic mode, it is useful to have it enabled. When enabled, it begins to collect
storage metrics that are otherwise unavailable or difficult to locate within vSphere.
This allows it to make more informed decisions if you decide to fully automate at a
later date. SDRS runs on a schedule that can be modified based on requirements. By
default, every 8 hours it analyzes the datastore cluster to evaluate whether any mi-
grations are required.
SDRS also provides affinity/anti-affinity rules for VMDKs; this means you can spec-
ify disk files that should reside together or apart. In general, it is best to leave them
together unless you have a specific reason to separate them.
Once enabled, many features are available to be configured for how to automate
your storage environment. Those options are as follows:
■ Free space threshold
■ Latency threshold
■ VM affinity/anti-affinity rules
The following figures demonstrate the available settings when configuring SDRS on
a datastore cluster. Figure 4-4 illustrates SDRS settings as displayed in the vSphere
web console.
SIOC
Storage I/O Control enables storage traffic shaping and priority within a vSphere
environment. This is feature similar to Network I/O Control (NIOC) and compute
resource pools. A congestion threshold can be enabled to guarantee storage traf-
fic priority. A latency threshold can be configured to ensure a datastore is always
achieving below a configured millisecond response time. If the datastore goes above
that number, vSphere attempts to limit the I/O from all hosts to that particular
datastore in an effort to lessen the latency. This setting can also be set to a dynamic
percent threshold to better calculate the appropriate millisecond response type de-
pendent on the type of storage used.
Shares can be assigned to VM disk files that can give them higher or lower prior-
ity in the event of storage contention. As with all shares, administrators must be
careful when assigning these values because incorrect settings can have a profound
impact on the environment. SIOC will also evaluate all VM disk share values to as-
sist in VM priority in times of congestion. Figure 4-5 shows the settings available for
SIOC.
Storage Policies
Storage policies are policies that can be applied to VMs to ensure the VM is placed
in the correct storage. These policies can be derived using vSphere Storage API
for Storage Awareness (VASA), which polls the storage device for its capabilities, or
custom profiles can be created. After these profiles are created, they are assigned to
datastores. When deploying VMs, the storage policy is selected to ensure they are
placed in appropriate datastores.
Storage policies are also leveraged by other VMware products such as vCloud
Director. vCloud Director uses storage policies to ensure vApps are placed in the
proper datastores at the vSphere level. This also enables tiering of storage services
within a cloud environment. Figure 4-6 demonstrates selecting a storage policy dur-
ing VM creation to ensure it is placed in the correct datastore.
Virtual SAN
A virtual storage area network is an object-based storage solution that aggregates
local disks into a policy-driven storage cluster. It is much more robust than the VSA
and offers support for up to 32 hosts in a cluster. Multiple types of storage profile
can be created to ensure performance that scales with a cluster. Multiple redundancy
options and cache percentages can be configured to provide very specific require-
ments to the storage platform. These profiles are then applied to VMs to ensure
they are placed in adequate storage for their requirements.
VSAN uses a combination of SSD and SATA/SAS hard drives to offer high per-
formance as well as capacity. VSAN is compatible with all vSphere features such as
vMotion, HA, and DRS. This product enables the functionality of a robust shared
storage platform utilizing the local capacity on the compute platform.
Virtual Volumes
Virtual Volumes (VVOLs) is a technology that extends VM-centric storage to stor-
age devices. Currently, a storage administrator creates a LUN or volume on a stor-
age device and presents it to a host, which then consumes this as a datastore. After
the datastore is created, VMs are placed on it. VVOLs changes this to a different
process: the virtualization administrator creates a VM, and this sends a command
to the storage device to create the VM’s disk file. Included in these commands are
requirements for the capacity and performance of the disk. The storage device then
creates and places the disk according to the requirements. This greatly simplifies
storage management and is another way to ensure that the VMs policies are adhered
to.
Availability
Availability of storage is of paramount importance in a virtual datacenter. When
storage is offline, the VM is offline. In circumstances such as a host failure or net-
work device failure, the guest VM can in most cases be brought back online in very
short order. To minimize the impact of device or cable failure, vSphere enables
multiple paths to shared storage devices. A storage device failure has a much larger
impact because all VMs on that device will be unavailable until the storage device is
brought back online.
VMware has many technologies that can minimize the effect of storage failures and
storage maintenance. Storage vMotion enables VM disks to be moved from one
datastore to another while the VM is online. This can occur between VMFS and
NFS datastores regardless of the backend storage type. It truly frees the VM from a
physical storage device. Enhanced vMotion is a new vSphere feature available in the
web client that allows you to change a VM’s host and datastore at the same time. It is
vMotion and Storage vMotion combined into a single operation.
Products such as VMware Virtual SAN and the VSA are designed to make local
storage devices as redundant as shared storage devices. These technologies replicate
the data from local datastores across multiple hosts. This enables servers with only
local disks to provide a highly available platform for running VMs. Figure 4-7 dem-
onstrates the possible failures that can be sustained and still maintain data integrity
and VM uptime.
SAN Switches
Disaster Recovery
Disasters occur and take many forms—a datacenter fire, data corruption, or even ad-
ministrator error. Ensuring the integrity and availability of VM data is of paramount
Virtual 1
machine
goes down X
VM VM VM VM VM VM
Restore
VDP
VDP
2 Select VM 3 Restore
Images/ Virtual Machine
Files to Recover
vSphere Replication
vSphere Replication (VR) is a vSphere feature that copies VM information from
one location to another. These locations can be within the same cluster, a different
cluster, a different vCenter, or even a different datacenter. It is a continual copy of
a VMs state performed on a defined interval. As with VADP, it replicates only the
changes made since the last interval to reduce network bandwidth and increase copy
times. Administrators can choose up to a maximum of 24 points in time to be able to
recover to; this can be configured as 1 a day for 24 days or 1 an hour for 24 hours.
This is dependent on your requirements for recovery. vSphere replication is a per-
VM technology and can back up a maximum of 500 VMs. The number of VMs that
can be backed up might be significantly lower depending on your available network
bandwidth and storage capacity.
vSphere Replication operates at the hypervisor level and is the only true hypervisor-
level replication technology today. VR via the hypervisor copies VM data from the
host to a specified appliance in another location. The appliance caches the data until
all the data has been received. It then applies this to the VM disks of the protected
copy at the remote site.
In the event of a failure at the primary site, the VM’s protected copy can be recov-
ered in the remote virtual center via the vSphere web client. This cannot be per-
formed if the original VM is still accessible. This is a per-VM activity. VMware has
a product called Site Recovery Manager (SRM) that can automate the recovery of
VMs in the event of a disaster. SRM is an additional purchase and is not included
with vSphere Replication.
Figure 4-9 demonstrates vSphere replication between vSphere hosts.
Source VM Protected VM
VM VM
VR
vSphere Replication Appliance
Summary
In this chapter, we learned about the differences between physical and virtual stor-
age. Physical storage is required to store the actual data, while virtualization ab-
stracts this storage to make it highly available and easy to consume. Storage devices
come in many forms; virtualization simplifies this and greatly reduces management
complexity. Virtualization reduces the impact of physical storage maintenance on an
environment by being able to migrate workloads away from certain storage devices
without interruption of service.
The concept of datastores and how they are accessed was detailed. Datastores are
logical containers of either NFS- or VMFS-based storage. These can be assigned to
datastore clusters, which simplifies management. Several protocols are supported,
with each offering its own features. The selection of protocol is dependent on the
overall environment design. vSphere works well with all the supported options.
Storage DRS and SIOC showed how vSphere simplifies management and ensures
that VMs get the storage resources they require. Storage policies can be applied to
VMs to ensure their placement in appropriate datastores. These features combined
ensure the performance of VMs is within established thresholds.
Other VMware storage technologies were highlighted and their features explained.
VMware has a robust suite of storage technologies for all sizes of organizations and
required business use cases.
Disaster recovery options were examined; vSphere Data Protection and vSphere
replication offer robust disaster recovery capabilities and ease backup management
tasks. VDP and VR utilize hypervisor-level technologies to minimize the impact of
backup processes and speed up recovery times in the event of a disaster.
4. Which is a virtual machine disk file created by VMware used to store virtual
machine data?
a. VFFS
b. VMDK
c. RDM
d. VMFS
6. Which feature limits host I/O to ensure datastore latency does not exceed a
specified threshold?
a. SIOC
b. NIOC
c. SDRS
d. DRS
8. What is the maximum number of intervals that can be stored by vSphere Data
Protection?
a. 24
b. 8
c. 16
d. 32
Applications within a data center must communicate with each other and with
users in varying locations; the medium used for this communication is the
network. Devices, whether physical or virtual, must communicate with each
other in a standard way that all devices understand. All virtual network devices
must be able to communicate seamlessly on the network and do so using vir-
tual switches. There are multiple types and configurations of virtual switches;
this chapter covers these types and configurations to provide the foundational
knowledge required to understand vSphere networking.
Table 5-1 “Do I Know This Already?” Foundation Topics Section-to-Question Mapping
1. Which is a feature that a physical network device can perform that vSphere
cannot currently perform?
a. Load balancing
b. Switching
c. Routing
d. Private VLANs
6. Which feature on a distributed virtual switch allows you to connect ESXi hosts
to physical switches by using dynamic link aggregation?
a. Load balancing
b. LACP
c. Active/Active adapters
d. NetFlow
Foundation Topics
Virtual switches and the physical network must be designed in concert with each
other to ensure the proper functioning of communication between devices. Redun-
dancy, availability, and performance must be architected into the environment at
every level to ensure the network delivers what is required of it. Figure 5-1 shows an
example of a logical design for connecting the virtual network to the physical net-
work.
Network Core/
Router
Physical
Switch
vSphere
Host
Virtual
Switch
Virtual
Machines
Each switch type has differing feature sets and use cases, as well as different cost and
skills required to administer.
NOTE Load balancing must work in concert with the physical network topology.
Incorrect load-balancing settings at the vSphere layer can cause intermittent traffic
issues.
NOTE Third-party switch knowledge is not required for the VCA-DCV exam; it
is included for informational purposes.
Portgroup
A portgroup is a construct created on the virtual switch that determines the network
connectivity for the attached VM. These portgroups are created per virtual switch
in the environment. A portgroup is a profile defining how a VM connects to the
network. These portgroups are then assigned to the VM’s virtual network card that
■ Connection settings
■ Network label—This name must match its equivalent portgroups on
other vSphere hosts.
■ VLAN ID (Optional)—If you are using VLAN trunking at the physical
uplink level, you must specify the VLAN ID number to tag on to traffic
for this portgroup. If you have a flat network without VLANs or specific
uplinks per VLAN, this setting is left as None.
Figure 5-5 shows a DVS portgroup and its settings in the vSphere web client.
■ VLAN information
■ None—No traffic will be VLAN tagged.
■ VLAN—Allows you to specify a VLAN ID for the network traffic.
■ VLAN trunking—Enables you to specify a VLAN range for the traffic
on this portgroup.
■ Private—Allows you to configure private VLANs on the DVS.
■ Teaming and failover
■ Load balancing
■ Route based on IP hash—Route based on IP hash uses a calcula-
tion based on the IP address of the traffic to select an uplink. When
this path is selected, it will stay until an event occurs that forces a
recalculation of the path.
■ Route based on source MAC hash—Route based on MAC hash
uses a calculation based on the source MAC address of the traffic
to select an uplink. When this path is selected, it will stay until an
event occurs that forces a recalculation of the path.
■ Route based on originating virtual port—Route based on the
originating virtual port uses a calculation based on the original
virtual port of the traffic to select an uplink. When this path is
selected, it will stay until an event occurs that forces a recalculation
of the path.
■ Use explicit failover order—Explicit failover uses the speci-
fied active and standby adapters. Traffic will flow over and active
adapter until that adapter becomes unavailable; traffic will then fail
to the next available adapter.
■ Route based on physical NIC load—This load-balancing option
evaluates the load on all uplinks for the portgroup. If one link is
over 75% and used for a 30-second duration, vSphere begins to
move traffic to a lower-used link.
NOTE Load balancing must work in concert with the physical network topology.
Mismatched load-balancing settings at either the physical switch or the vSphere
layer can cause intermittent traffic issues.
There are other uses for vmk adapters such as accessing network-attached storage
and for VMware fault tolerance (FT). You should follow VMware and vendor best
practices for the configuration of particular vmk adapter types. Figure 5-6 shows the
vmk adapter options in the vSphere web client.
Custom share values and profiles also can be created; therefore, care must be taken
when using custom share values to ensure priorities are properly assigned. Custom
share values can be complex and introduce risk into the environment. Custom pro-
files assigned to portgroups will override the default NIOC profiles for that type of
traffic.
As with compute resource pools, NIOC priorities apply only in periods of conten-
tion. That is to say, that unless the host is constrained for network resources, all pro-
files will be treated equally. When contention occurs, machines with high priority
will be given greater access to network resources than those with normal priority.
All limits applied to network profiles will always be in place. It is highly recom-
mended to never use a limit unless you have an extremely good reason to do so.
Incorrectly implemented limits can have a very detrimental effect on the proper
functioning of applications. Figure 5-7 displays NIOC settings in the vSphere web
client.
Quality of Service
Quality of service is a way of prioritizing network traffic for a network environ-
ment. Whereas NIOC can control traffic only when it enters and leaves the vSphere
environment, QoS tags can function throughout an entire network environment as
long as it is configured end-to-end on all devices in the network. The QoS value is
configured on the same network profile NIOC uses but is not honored by vSphere.
vSphere uses the network share value and merely passes the tag along with the traf-
fic to the upstream infrastructure. QoS coupled with NIOC enables end-to-end pri-
oritization of network traffic throughout the datacenter.
QoS works on a rating scale with 0 at the highest and 7 at the lowest, which means
that traffic tagged with a 1 will have priority of traffic tagged with a 6. This is useful
in environments with applications that require very low latency, such as voice com-
munications.
NetFlow
NetFlow is an industry-standard feature that provides the ability to collect network
traffic that ingresses or egresses a network interface. Network flow information can
be sent from a DVS to a designated collector in the environment. This collector
can then view valuable network information, such as the source and destination of
traffic, and determine potential network bottlenecks and which protocols are in use.
This information is invaluable in network troubleshooting.
Summary
In this chapter, we learned about physical networking and how it differs from virtual
networking. Virtualization abstracts the physical network, simplifying network man-
agement and providing high availability. Virtualization is expanding more and more
into the physical network realm.
We discovered the concept of virtual switches and their features. The differences be-
tween a vSphere standard switch and distributed virtual switch were explored. Both
switches use portgroups to define network configurations that are then applied to
VM network adapters. Network traffic flows from virtual network adapters through
the virtual switches physical uplinks to the physical network, and vice versa.
Network I/O control and quality of service ensure VMs get the access to network
bandwidth that they require. NetFlow enables networking monitoring devices to
gather information about virtual network devices, offering end-to-end visibility.
2. Which are two types of virtual switches available with vSphere? (Select two.)
a. Logical standard switch
b. Consolidated virtual switch
c. Distributed virtual switch
d. vSphere standard switch
6. Which feature on a distributed virtual switch allows you to tag traffic priority
in a manner respected by physical switches?
a. Quality of service (QoS)
b. Network I/O control (NIOC)
c. Physical load-based teaming
d. NetFlow
7. How many physical uplinks at minimum are required for beacon probing?
a. 1
b. 2
c. 3
d. 4
10. Which is a construct that defines the connection parameters for VMs on a
virtual switch?
a. Portgroup
b. Network profile
c. vSwitch
d. vmk
Table 6-1 “Do I Know This Already?” Foundation Topics Section-to-Question Mapping
7. Which of the following technologies can analyze the environment for optimi-
zation challenges?
a. vSphere HA
b. vCenter Server Operations Manager
c. VMware vMotion
d. vSphere Replication
8. Which optimization technology enables you to allocate more disk space than
is available in the host system?
a. Templates
b. Fast provisioning
c. Dynamic Resource Scheduling
d. Thin provisioning
Foundation Topics
Hardware Availability
To solve common hardware availability problems VMware offers the following solu-
tions:
■ “My company needs to minimize the downtime caused by a host failing in the environ-
ment.”—vSphere HA will restart VMs in the case that there is a failed host in
the environment, therefore minimizing downtime.
■ “My company has a service level agreement that calls for a very quick response to a
host failing. Therefore, I have to be sure that VMs are up and running as quickly as
possible after a failure.”—vSphere HA automatically monitors for network con-
nectivity and datastore heartbeats. As a result, as soon as a failure occurs, VMs
will be failed over.
■ “I need to ensure that if a host loses access temporarily, appropriate action is taken.”—
vSphere HA monitors network availability for isolations and acts only if a host
is truly isolated and configured to failover in that condition.
Application Availability
To solve common application availability problems, VMware offers the following
solutions:
■ “We have an application that cannot be interrupted by hardware failures.”—vSphere
FT creates a shadow VM and will immediately switch over in the case that a
host failure is detected to ensure that there is no downtime for the application.
■ “I need to ensure that if my application stops responding, action is automatically taken
to try to resolve the problem.”—vSphere App HA monitors common business ap-
plications using VMware Hyperic application monitoring. It detects and cor-
rects problems with the service by restarting the service or server as configured
in the policy created by the administrator.
■ “My company requires that if an operating system stops responding for a period of
time, action is automatically taken to try to resolve the problem.”—vSphere HA –
VM Monitoring uses VMware Tools heartbeats to validate whether the guest
OS is running. If a heartbeat is not received within a defined period, the VM
will be restarted.
Maintenance
To solve common maintenance problems, VMware offers the following solutions:
■ “I need to be able to perform hardware maintenance at any time without impacting
the availability of services.”—vSphere vMotion enables VMs to be live migrated
to different ESXi hosts to allow hosts to be placed into maintenance mode and
taken offline.
■ “We need to avoid taking multiple servers down to apply hotfixes.”—vSphere vMo-
tion enables VMs to be live migrated to different ESXi hosts to allow hosts to
be placed into maintenance mode.
■ “My company needs to perform hardware updates at any time of the day without
impacting users who are working.”—vSphere vMotion enables VMs to be live
migrated to different ESXi hosts to allow hosts to be placed into maintenance
mode.
Environmental Availability
To solve common environmental availability problems, VMware offers the following
solutions:
■ “Our process to back up servers is difficult. It is hard to test for validity to ensure that
they are recoverable, and it is hard to restore an entire server quickly.”—VMware
Data Protection enables backups to be taken of entire VMs, which are easy to
test and recoverable very quickly.
■ “My company has mission-critical applications that must be available even if environ-
mental conditions cause problems at the datacenter. We would like a solution for this
that will also enable us to easily test the environment to ensure that any action that
is taken will be successful if we do have a disaster.”—VMware Site Recovery Man-
ager allows an administrator to configure a plan to failover an entire site to a
backup location. It accomplishes this by using backend array replication (or
vSphere Replication) to replicate changes that are applied based on the config-
ured recovery point and recovery time objectives.
■ “We would like to replicate server data to another site but have different types of
storage array at each site.”—vSphere Replication is a technology built in to the
ESXi architecture that enables VMs to be migrated between datastores regard-
less of the underlying storage technology.
Datacenter Management
To solve common datacenter management problems VMware offers the following
solutions:
■ “I am the only administrator for the environment, and I need a centralized way to
manage my datacenter.”—vCenter Server enables administrators to get a cen-
tralized view of the entire environment after it has been virtualized and allows
vSphere solutions such as high availability (HA) to be used.
■ “My datacenter is almost out of space. Power and cooling are maxed out, and it is diffi-
cult to introduce new hardware into the environment.”—VMs enable physical hosts
to be consolidated so that a single physical server can act as many, sharing the
resources of the server and consuming less physical space, power, and cooling
capacity in the datacenter.
■ “Our environment is complex, and we have many administrators around the world.
We need to be able to see everything around the world in a single pane of glass for
administrators.”—vCenter Server enables administrators to get a single-pane-
of-glass view of their virtualized environment. This includes being able to link
multiple vCenter servers together.
Provisioning Management
To solve common provisioning management problems, VMware offers the following
solutions:
■ “We have many physical servers running on really old hardware. This hardware is
out of warranty, and the applications are difficult to reinstall.”—Virtual Machines
and VMware Converter can be used to convert the older hosts into VMs, thus
keeping the applications available, even if they are legacy applications.
■ “I need to be able to provision a server from a standardized image that my organiza-
tion uses.”—vCenter Server can be used to create standardized templates that
can be used for deployment of resources.
■ “My company is growing quickly, and we need a way to rapidly deploy many VMs to
meet the demand.”—vCenter Server templates can be used to deploy many VMs
rapidly.
Compliance Management
To solve common compliance management problems, VMware offers the following
solutions:
■ “My company is in retail and needs to ensure we meet compliance regulations for credit
card information.”—vCenter Configuration Manager can be used to check com-
pliance of any regulation that has been configured and comes preloaded with
common baselines.
■ “We need to monitor our environment to make certain that our servers are running
the latest patches. If they are not, an administrator needs to be notified.”—vCenter
Configuration Manager can be used to check whether servers have specific
patches installed on them.
■ “I need to monitor an environment for unexpected changes to it. This will help to pre-
vent change control from being bypassed for critical servers”—vCenter Configura-
tion Manager can be used to check servers on a regular cadence to ensure that
they meet specific standards.
Scalability Challenges
Virtualization solves many of the scalability challenges mentioned in Chapter 2 in
unique ways. VMware provides solutions for all these challenges given the various
products in the vCloud suite. In this section, we discuss the solutions to these chal-
lenges.
Infrastructure Scalability
To solve common infrastructure scalability problems, VMware offers the following
solutions:
■ “We are a small company and do not have a SAN. We still want to be able to take ad-
vantage of virtualization.”—VSA or VSAN will enable local storage to be used
and presented to the hosts as shared storage, so there is no requirement for
shared storage to take advantage of the benefits of virtualization.
■ “My company has the need to scale out the network bandwidth available to my hosts
without taking downtime on the VMs.”—DVSes enable additional network up-
links to be added, so that additional bandwidth can be provided to a switch.
VMware vMotion and DRS allow hosts to be evacuated and taken offline to
install additional NICs.
■ “I need to ensure that I can control the characteristics of the storage that is used to en-
sure that I have enough room to scale my environment appropriately to meet the needs
of my servers.”—VSAN enables the performance characteristics of the VM stor-
age to be configured dynamically with no downtime on the VMs or migration
to a different storage device.
Optimization Challenges
Virtualization solves many of the optimization challenges mentioned in Chapter 2 in
unique ways. VMware provides solutions for all of these challenges given the various
products in the vCloud suite. In this section, we discuss the solutions to these chal-
lenges.
Monitoring Optimization
To solve common monitoring optimization problems, VMware offers the following
solutions:
■ “We would like to be able to proactively monitor for performance bottlenecks in our
environment. Currently our environmental monitoring provides only a few basic
statistics that we can look at.”—vCenter Operations Manager enables an admin-
istrator to proactively monitor environment for bottlenecks and other perfor-
mance problems.
■ “My company needs to have detailed performance reports that can be used to trend our
environment.”—vCenter Operations Manager provides detailed performance
reports and heat maps of an environment to be analyzed by an administrator
or stored for historical purposes.
■ “I have several servers that are idle most of the time. I need a way to monitor them
and be able to assess whether consolidating them will affect performance.”—vCenter
Operations Manager provides projections of performance on a system and as a
result can predict what utilization will be on a system in the future given con-
straints to the resources.
Efficiency Optimization
To solve common efficiency optimization problems, VMware offers the following
solutions:
■ “We have several servers that have very little disk usage but were configured with a
large amount of disk space. It would be great to be able to minimize the impact this has
on the overall disk usage.”—Thin provisioning enables servers to not fully con-
sume the full physical storage space that has been allocated to the VM until it
is actually used for preserving disk space. These disks will grow as needed (up
to the maximum configured size), and an administrator therefore needs to be
vigilant about monitoring free space to ensure that errors don’t occur due to
low disk space.
■ “I need a way to detect whether I am efficiently using the available resources in my
environment.”—vCenter Operations Manager can detect whether resources are
being used efficiently in the environment.
■ “My company has several storage arrays that are used in the environment. Some of
these storage arrays are high performance and cost a lot more to operate than oth-
ers. Many times development users place VMs on the incorrect storage device, causing
degradation of performance for the array, which we correct by moving the VM to the
right class of storage. We need an automatic way to ensure that users can choose the ap-
propriate class of storage from the time of deployment.”—Storage profiles enable an
administrator to configure classes of storage. Thus, initial placement of VMs
can be ensured to be on the proper class of storage.
Performance Optimization
To solve common performance optimization problems, VMware offers the following
solutions:
■ “I need a way to proactively ensure that I have a healthy environment so I can get
optimal performance of my hosts.”—vCenter Operations Manager enables an ad-
ministrator to assess the resource utilization on hosts to ensure they are oper-
ating efficiently.
■ “We need a way to prioritize network traffic to ensure that our mission-critical VMs
get the peak performance possible.”—QoS for DVSes, and network I/O control
enables traffic to be prioritized for optimal performance of certain types of
traffic.
■ “My company needs to ensure that we can add resources to VMs in our environment
on-the-fly to address performance problems if they occur.”—Hot add resources can
be added to VMs to ensure that adequate resources are available to the server
to process incoming requests.
Regardless of the technologies being used, these are the types of challenges that can
impact the design based on the size of the business. Enterprises can often have very
unique requirements that lead to larger expenses or a completely different solution
being required.
Summary
In this chapter, we discussed how the different virtualization technologies VMware
offers solve the challenges an administrator faces when designing and implementing
his environment. Technologies that meet the availability, management, scale, and op-
timization challenges of the environment were discussed.
Each VMware technology has a unique place in an environment that will revolution-
ize not only the way you operate in the environment, but also the way that you think
about the common business challenges.
Key Technologies
List the technologies that solve the challenges for the Key Topics from this chapter,
and check your answers in Appendix A.
Availability Challenges, Management Challenges, Scalability Challenges,
Optimization Challenges
Review Questions
The review questions section is for you to review your knowledge of the topics in
the chapter. You can find the answers in Appendix A.
1. Which of the following technologies will solve availability challenges? (Choose
all that apply.)
a. VMware FT
b. vCenter Operations Manager
c. VMware HA
d. Dynamic Resource Scheduling
Chapter 1
2. B
3. A, B, C
4. C
5. A, D
6. B, D
7. A
Chapter 2
2. B
3. B
4. A
5. A
6. A
7. D
8. B
9. C, D
10. A
Chapter 3
2. B
3. C
4. B, C, D
5. B
6. A, D
7. D
8. A
9. B
10. B
11. A, B, C
12. B
13. B, C
14. A
15. A, B, C
16. A, B
5. A and C. VMware HA and VMware DRS are the only resources for which a
cluster is required. Resource pools can be configured on hosts that are not a
part of a cluster. vCenter Operations Manager also can monitor hosts that are
not a part of a cluster.
6. C and D. Resource pools enable shares and reservations to be configured for
CPU and memory resources. They do not enable the configuration of HA or
CPU affinity settings.
7. B and D. DPM uses vMotion to migrate the running VMs to different hosts;
then the host is put into standby mode, thereby preserving power in the envi-
ronment.
8. A. Fault tolerance provides continuous availability to VMs by replicating the
state of all activities on a VM to a secondary server. If a failure occurs, the sec-
ondary VM immediately takes over processing for that VM.
9. B and C. Application HA can either restart the failed service and/or restart the
guest operating system. It will not restart any ESXi host or vCenter services in
response to a service failure.
10. B. VMware Data Protection provides capabilities to quickly and efficiently
back up and restore VMs. It does not provide abilities to replicate to the cloud,
malware protection, or configuration management of the servers.
11. B and C. VMware SRM is used to protect the datacenter from a failure or
condition at a single site. Thus, if there is a flood or air conditioning failure
that is impacting an entire site, SRM can be used to failover mission-critical
servers to keep the business up and running. Power failure of a single host or
installation of new hosts would not require utilization of SRM.
12. B. The minimum recovery point objective for vSphere replication is 15 min-
utes. Any lower than this, as of vSphere 5.5, is not currently supported.
13. D. VMware vMotion is used for live migration of VMs between ESXi hosts.
None of the other technologies mentioned use VMotion.
14. B. vCenter Operations Manager provides the efficiency metric to allow ad-
ministrators to monitor for how efficiently the resources are being used. This
enables the administrator to proactively assess the environment to ensure re-
sources are not being wasted.
15. D. VMware Configuration Manager provides compliance management for the
environment, which can later be applied to ensure that all company or work-
load standards (ex. PCI DSS) are monitored and met.
Chapter 4
2. A
3. A
4. B
5. D
6. C
7. A
8. C
9. A
10. A
11. B
12. C
Chapter 5
2. C
3. D
4. A
5. B, C
6. B
7. B
8. D
Chapter 6
2. C
3. B
4. A, B
5. D
6. A, C
7. B
8. D
LACP The Link Aggregation Control Protocol is a network specification that en-
ables the bundling of several physical ports together to form a single logical channel.
Load balancing A network feature that enables the intelligent placement of traffic
for optimal performance.
Management Management is the means by which all resources are administered.
Management challenges Challenges that relate to the management of the envi-
ronment. These challenges could be as simple as ensuring there is centralized man-
agement of resources to being something as complex as consolidation of datacenter
resources to allow for further expansion.
Management challenge solutions VMware vCenter Server, VMotion, Storage
VMotion, templates, clones, vSphere Configuration Manager, and virtual machines.
NAS Network attached storage is a storage device connected to vSphere via Eth-
ernet and presents file-based storage.
NetFlow NetFlow is an industry-standard feature that provides the ability to col-
lect network traffic that ingresses or egresses a network interface.
Network A network is the means by which devices such as switches, routers, and
network adapters connect and communicate with the datacenter and outside world.
NFS Network File System is an industry-standard distributed filesystem protocol
available via Ethernet.
Optimization Optimization is making the most efficient use of available re-
sources.
Optimization challenges Challenges that relate to the effective utilization of the
resources in the datacenter. These challenges relate to monitoring the environment
for wasted space or to ensure that you are seeing peak performance of the resources.
Optimization challenge solutions vCenter Operations Manager, Dynamic Re-
source Scheduling (DRS), thin provisioning, storage profiles, and QoS for DVSes.
P2V migrations A physical server converted to a virtual machine.
Portgroup A portgroup is a profile defining how a VM connects to the network.
RDM A raw disk mapping is a disk type that gives virtual machines access to raw
storage.
Resource pools These provide a method of consolidating and managing compute
(CPU and memory) resources to enable reservations, shares, and limits to be set.
SAN Storage array network is a block-based network for storage traffic.
vMotion Enables live migration of running virtual machines between ESXi hosts.
VMware Data Protection (VDP) Enables the backup and restore of virtual ma-
chines.
VMware SRM Enables an entire site to be failed over according to a recovery
plan, in the case where you have a problem that spans an entire rack or datacenter.
vSphere App HA Provides application service monitoring via the VMware Hy-
peric agent and restarts the service or even the virtual machine to try to correct an
error.
vSphere DPM Part of VMware DRS that monitors ESXi host power utilization
and will put hosts into standby (and consequently take them out of standby when
needed) to save power during non-peak periods of utilization.
vSphere DRS A technology that uses vMotion to live migrate resources between
hosts in a cluster to ensure that resource utilization is optimized for the virtual ma-
chines.
vSphere FT Provides continuous availability to configured virtual machines, al-
lowing for a shadow virtual machine to take over in the case that the primary fails.
vSphere HA Monitors the hosts and virtual machines for failures and restarts
virtual machines to other hosts if one is detected. It uses both the network and the
storage infrastructure to ensure that servers are truly failed before it acts.
vSphere Replication (VR) Replicates virtual machine data between ESXi hosts to
enable quick recovery of virtual machines.
* Available to new subscribers only. Discount applies to the Safari Library and is valid for first
12 consecutive monthly billing cycles. Safari Library is not available in all countries.
Mobile Apps
Pearson is the official publisher of Cisco Press, IBM Press,
VMware Press and is a Platinum CompTIA Publishing Partner— Newsletters
EXAM TIPS AND TRICKS from Pearson IT Certification’s Question of the Day
The VCA-DCV Official Cert Guide from VMware Press comes with three
practice exams. Each practice exam contains 50 questions, simulating the real exam,
and focuses on different objectives from the official VMware blueprint. Take each
practice exam in full and revisit the questions you answered incorrectly for remediation
and to identify the topic areas in which you need additional study support.
Follow the steps below to register this eBook and access the test questions:
1. Go to http://www.informit.com/register and log in or create a new account.
2. Enter this ISBN: 9780133854497
3. Answer the challenge question as proof of purchase.
4. Click on the “Access Bonus Content” link in the Registered Products section
of your account page to be taken to the page where your downloadable content
is available.