XenDesktop Bandwidth: The complete set
Part 1 - Prologue: Methodology and infrastructure
Part 2 - by the Numbers: Take the time to optimize
Part 3 - Bringing it All Together: average daily users and general recommendations
Part 4 - What About XenApp
part 5 - Do It Yourself: Starter Kit
part 7 - Who needs a part 6 when discussing XenDesktop 7.x
customer bonus (And XenApp 7.5)! Analysis
Bonus HDX 3D Pro
Part 2: By The Numbers: Take the time to optimize
In Part 1, I discussed the methodology and the infrastructure used for testing.
In this post I will discuss the policy settings and optimizations that we found to be most beneficial for WAN optimization deployments and why is essential for a WAN deployment.
We tested three different configurations:
- OOBE : A default "out of the box" configuration
- optimized : optimized WAN configuration using both Citrix policies and Windows registry changes that much impact on the look and feel of Office
- Pushing the Limit a highly optimized configuration that changes the appearance of the office, but requires less bandwidth
Before you start thinking: "Wait, why not use CloudBridge" the answer is that it would a good idea, but we have not completed testing at that time. Would it improve outcomes? Absolutely. Will it make your life easier? Very likely. Will I be tested in the future? Hopefully ! for now, see the white paper recently published about some of the benefits CloudBridge. More to come on CloudeBridge in part 3 as well.
Before You Begin
As I mentioned in Part 1 of this blog series I'll mention here the other, there are several warnings I have to do in terms testing. The first (which I mentioned in Part 1) is that tests with Login VSI not send mouse clicks and keystrokes from the client to the server and therefore the bandwidth of one side of the server is discussed here .
Second, a physical laptop was used as the endpoint operating device sessions at a resolution of 1440 × 00 (between 720 and 1080) for the configurations tested. The resolution of the session will affect the use of bandwidth. So if you plan to test, make sure to test a reasonable resolution comparable to what users experience.
Third, the results presented in this blog do not include the logon process. Indeed, one of logon process requires a significant amount of bandwidth for a short time and unrelated to the operation of applications. Graphic bandwidth logon is provided at the end of this post
Configuration 1 :. OOBE Default
Many customers VDI deployments use images based on physical image of companies; those with little or no optimization policies in place. Therefore, the first configuration is based on the deployment scenario. In a LAN environment, the bandwidth will not be a problem even with this kind of deployment, although some optimizations can also give better scalability in terms of CPU and memory and should be considered, even on the network local. However, when switching to a WAN environment, the bandwidth is an entirely different story.
An image of Windows 7 by default will include many styles and visual parameters that most users do not notice, but require a considerable amount of bandwidth. Many of these parameters are "eye candy" and I would say that they should always be turned off in a virtual environment. To access these settings, open the dialog "Advanced System Settings" and under "Performance", click the "Settings" button. What you should look at is something like the screenshot below (right shows a desktop computer compatible with Windows Aero with additional options).
Disabling these visual styles has long been considered a best practice, although in many deployments, they are left enabled. In the default configuration, only two of the above settings have been disabled; "Animate windows when minimizing and maximizing" and "Show shadows under mouse pointer". Different deployments may end with different settings enabled if not properly configured. Aero is not enabled in the default configuration (the default XenDesktop 5.6 policy is to disable Aero Redirection). Note that this is not the case in XenDesktop 7 where there have been significant improvements in the bandwidth required for the redirection Aero and it is enabled by default for Windows clients
Configuration 2 :. Optimized
For the second configuration we set up some of our own best practices Citrix. The goal was to create a configuration that reduces bandwidth (rather significant), but does not significantly alter the desktop experience. Most of these changes are nothing new and include disabling eye candy mentioned above. They were taken from the literature, such as the optimization guide and planning guide policy. There are some additional configuration changes other than those guides who were primarily specific application
The changes were implemented both through political and Citrix Windows registry changes . find details of these in a suite DIY blog. For now, here are some of the main bandwidth reduction policies that we have set.
- Citrix Policies
- Reduce maximum frames per second to 15 (default is 24)
- Enable additional color Compression
- animation Disable
- Disable Wallpaper
- Disable Show window contents while dragging
- Disable all settings of visual effects in performance options except for "use visual styles on windows and icons"
- Adobe to Optimize Terminal services uses under Article
- Clear start screen for Microsoft Word, Excel and PowerPoint
Warning: serious problems might occur if you modify the registry incorrectly. For added protection, back up the registry before you edit it. See more about backing up and editing the registry here .
Note There are several ways to implement the registry changes above. In this case, local profiles were used with a desktop image pooled and we decided to use Group Policy Preferences. Some changes had to be made to the defaults that are read HKLM when a new profile is created. Changes can also be made to the initial desktop image or HKCU. Results may vary based on your profile solution and office type, so I recommend you to test this extensively to ensure that all the changes took effect. I'll discuss more in the DIY blog.
The first results are very promising. The bandwidth has dropped significantly across the board for all applications, in many cases over 50%. The table below shows the average bandwidth of both the default and optimized configurations in a LAN environment.
However, these tests were performed on the local network where bandwidth is rarely a concern, so what about a T1 connection? The results were equally impressive with savings of up to 82% (see results summarized in the table below). Excel showed the strongest economies and was very sensitive to even lower bandwidth limits. PowerPoint presented the smallest savings due animations between slides.
savings with similar results have been regularly observed at all tested levels of bandwidth, although this blog will mainly discuss the results found at 1.536Mbps because of enormous amount of data collected. Reactivity measured in manual testing (discussed in the blog 1) also improved across the board for all applications with lower bandwidth limits and a table with all the results of manual testing can be found here.
Configuration 3: Pushing The Limit
final configuration is focused on further reducing bandwidth consumption with several simpler optimization. In this scenario, the look and feel of the office has been assigned, and the resources of the server because of the heavy compression. These configuration changes have allowed for savings of additional bandwidth and reactivity gains in manual testing. How the additional bandwidth savings has been achieved? For starters, all Citrix compression policies were set for the maximum setting. In addition, all visual styles that made the office look like the Windows 7 desktop has been disabled. These parameters are summarized below. However, these should be carefully considered that the user experience is inherently different.
- Citrix Policies (In addition to those of the second configuration)
- Reduce maximum frames per second to 10
- Lossy compression level: Low
- minimum Image quality: Low
- Heavyweight compression Enabled
- Audio quality: Low
- Disable all redirection (printers, drives, ports, USB, TWAIN)
- maximum allowed color depth: 16 bits per pixel
- Disable all settings of visual effects in the performance options
- Disable blinking cursor
These projected changes bandwidth savings that have not quite match the initial savings seen from the optimized configuration, but compared to the optimized image they were certainly important. The table below compares the configuration "maximum optimized" against the "optimized" configuration. The difference in terms of kbps recorded is significantly lower than among OOBE and optimized, but for a highly constrained environment of small differences can have a significant impact for users. Because these changes really affect the user experience, not implement them unless you really need. first use of the bandwidth test strip to avoid unnecessarily degrade the user experience.
Graphics Charts
I myself am a visual person and I prefer to digest my graphic content. I believe many others are so below I are different graphical presentations of our findings in deep dive application. Select the thumbnails to see the images.
Next Time
in part 3, I will discuss the average observed during the execution of the average workload Login VSI which includes Internet browsing, video 480p, and the idle time to simulate the behavior typical of the user. I will discuss the "daily average" and general recommendations based on the tests.
Thanks for reading,
Amit Ben-Chanoch
Worldwide Consulting
Office applications team
Accelerator project
Virtual Desktop Manual
Follow @CTXConsulting
0 Komentar