Book Image

Optimizing Citrix?? XenDesktop?? for High Performance

By : Craig Thomas Ellrod
Book Image

Optimizing Citrix?? XenDesktop?? for High Performance

By: Craig Thomas Ellrod

Overview of this book

Citrix XenDesktop is a suite of desktop virtualization tools designed to provide users with fast and convenient access to their Windows desktops and applications through any device. Virtual desktops mean that rather than setting up hundreds or thousands of individual computers in an enterprise, companies can instead opt to create servers with large amounts of memory, disk, and processing resources, and use virtualization to offer these resources to end users. The result of this is that users are provided with an experience that appears to be identical to having an individual desktop PC. Each user has some disk space, processor time, and memory allocated to them, as though it is present on their own physical machine, when in reality, the resources are physically present on a centralized server. This book starts by answering the basic questions you need to ask when considering XenDesktop, followed by methods of how you can properly size your server infrastructure for XenDesktop. You’ll discover how to optimize the virtual machines used in XenDesktop, how to optimize your network for XenDesktop, and how to optimize the hypervisor and the cloud. You’ll also learn how to monitor XenDesktop to maximize performance. By the end of the book, you will be able to plan, design, build, and deploy high performance XenDesktop Virtualization systems in enterprises. You will also know how to monitor and maintain your systems to ensure smooth operation.
Table of Contents (18 chapters)
Optimizing Citrix XenDesktop for High Performance
Notice
Credits
About the Author
About the Reviewers
www.PacktPub.com
Preface
Index

Getting started with hypervisors


Before we get started, you will need to understand what a hypervisor is. A hypervisor is a thin operating system that hosts multiple instances of disparate operating systems. It can also be defined as software that can create and run virtual machines. The hypervisor software runs on server hardware that is enabled for virtualization. Once this is installed, you can install several instances of different operating systems onto the hypervisor. The hypervisor was game-changing, because instead of running one operating system per server, you could now run X number of operating systems on one server, saving space and money.

There are several vendors that make hypervisors, such as Citrix XenServer, VMware ESXi, Microsoft Hyper-V, and the Linux open source Kernel Virtual Machine (KVM). There are two types of hypervisors:

  • Type 1 hypervisors, which run directly on the server hardware. These are also known as "bare-metal" hypervisors.

  • Type 2 hypervisors, which run on top of an operating system, which then run on the server.

As you can imagine, the Type 1 hypervisors have been touted to have better performance as they interact directly with the server hardware resources.

Citrix XenServer and VMware ESXi are Type 1 hypervisors. Microsoft Hyper-V is presumably a Type 1 hypervisor. There has been debate over whether Hyper-V is a Type 1 or Type 2 hypervisor mainly because you first install the Windows server operating system and then turn on the Hyper-V role, giving the perception that Hyper-V is running on top of or with the help of the Windows Server operating system. Obviously, XenDesktop runs on Citrix XenServer, but it can also run on VMware ESXi and Microsoft Hyper-V.

Note

At the time of this writing, XenDesktop does not run on KVM. You could probably make it work. Citrix does not officially support KVM. Tribal Knowledge says you could run XenDesktop on KVM, but you would not be able to use MCS or PVS automation for creating VMs.

The following diagram gives you a visual idea of the differences between the types of hypervisors compared to traditional servers. It also shows how the interaction between these components contends for hardware resources and ultimately affects performance and sizing of hardware resources: