VMware View 5.0 User Experience Enhancements

VMware View 5.0 is all about user experience enhancements. VMware finally included RTO profiles into the product, and in conjunction with Teradici made several changes to PCoIP display protocol. Those changes aim to provide better user experience for users over slow links.

PCoIP display protocol is known for providing host based lossless imaging using multiple encoding algorithms at any given time. Those algorithms are applied to different display content as required. PCoIP scans the display image at the frame buffer and compresses different content within the frame with the appropriate encoder. Because PCoIP captures the entire frame buffer and sends it to the end-point device for decompress the image is always exactly the same to the original.

Using the approach mentioned above PCoIP guarantees the image quality; or what Teradici likes to call “Pixel Perfect”. However, because of that PCoIP always tax networks requiring additional bandwidth to finalize the “Pixel Perfect” transmission.

Teradici and VMware made enhancements to the display protocol in three different areas. According to VMware the testing results demonstrate up to 75% bandwidth reduction when compared to previous releases of PCoIP and VMware’s historical data.


Lossless CODEC

The first area is the lossless CODEC. Cleartype / Anti-alias fonts account for about 24% bandwidth with most remoting protocols. Changes were made to the lossless CODEC to improve this. This is On by default and you do not need to do anything.


Client Side Caching

The second area was the implementation of client side caching. Caching is turned On by default. Currently, client side caching does not work with Zero clients or the VMware View mobile clients like iPad/Android. It works with all our other clients though, and VMware is promising to deliver a Mac client with client side caching support later this year.

VMware View Client uses image caching to store portions of the display that were previously transmitted. Image caching helps to reduce the amount of data that is retransmitted.  When client side caching setting is not configured or disabled, a default client image cache size of 250MB is created and used.

When client side caching is enabled, it is possible configure a client image cache size from a minimum of 50 MB to a maximum of 300 MB. The default value is 250MB.


Build to Lossless

The third area is the ability to disable build to lossless. Here there are three stages. Initial image, Perceptually lossless and lossless. By disabling BTL it is possible to deliver a perceptually lossless experience which is adequate for the vast majority of users. This results in lower bandwidth usage.

The lossless experience is the default mode so you need to disable it if you want the bandwidth saving vs. image quality. As per previous iterations if the protocol this is a Group Policy setting that is imported into Active Directory from the PCoIP.adm file template.

Results will vary based on the video workload.

According to VMware, using their standard View Planner as benchmark they see the following benefits:

  • View 4.x with default settings = 150Kbps
  • View 5.x with default settings = 60Kbps
  • View 5.x with BTL off = 48Kbps

It is important to understand the use case, application image and bandwidth requirements before deciding to disable BTL.

Lossless is the default setting and one that will deliver the richest end user experience. The default time to build to lossless is 30 seconds. End users with high requirements for image quality are great candidates for this setting. Additionally any users on a LAN environment benefit from a fully lossless experience.

Perceptually Lossless (BTL Disabled)

Typical task or knowledge workers are good candidates to deliver a perceptually lossless experience as standard productivity applications are not impacted by image quality. Users in high latency, low bandwidth environments benefit from this setting as it reduces the bandwidth used per session, increases the user density and scalability per WAN link.

Leave a Reply