• Technical IT

    Solutions delivered throughout the UK
  • Business Applications

    Solutions delivered throughout the UK
  • Professional Best Practice

    Solutions delivered throughout the UK
  • Professional Development

    Solutions delivered throughout the UK

VMware View & vSphere – Performance & Best Practices

21 November 2011

Performance Improvements in View 5:

PCoIP protocol improvements
Optimizations to the PCoIP protocol benefit users in almost all usage scenarios.

Bandwidth Optimization
VMware View 5 offers major efficiency improvements in the WAN environment where bandwidth is highly shared. This is especially impressive as it does so whilst maintaining the user experience. In a typical office user, bandwidth consumption is reduced by up to 75%. This significant improvement means the VDI can support 4x more users to the same network. Even with new caching and lossless codec features, overall guest CPU usage is reduced.

Compute Optimization
Idle CPU usage is improved. Whereas once when a screen had loaded, a machine can lie dormant for varying periods of time whilst a user reads a page, VMware View 5 reduces the amount of overhead background activity and bandwidth used in this ’downtime’. Image processing and compressor functions have been reviewed, their coding altered for efficiency and implementation further optimized.
With View 5, PCoIP Protocol has been optimized to take greater advantage of the SSE SIMD support. This SIMD acceleration not only provides an expanding coverage of the code base but also applies the SSE instructions on the latest processors.*


Platform improvements with VMware vSphere 5
CPU scheduler
CPU and memory optimizations in vSphere 5 help achieve a higher consolidation ratio, particularly when using Windows 7 as this uses more memory than Windows XP.
VMware vSphere 5 has scheduler optimizations and works by making efficient use of the CPU. It does this by optimally scheduling the different desktops on the hyper-threaded core, meaning all thin (and zero) clients run in parallel whilst performing multiple tasks.

Memory footprint reduction
Memory on demand and swap mechanism lets more of the guest VRAM stay in the main memory. 60% footprint reduction = better consolidation rate.

Conservative ballooning
In earlier versions, the balloon driver activated when the systems free memory dropped below 6% of the total memory. On servers with a large amount of main memory, 6% is a lot and ballooning is not required. vSphere 5 keeps the 6% threshold for machines with less memory but also allows ballooning at as low as 1% for larger machine memories.

vSphere 5 also reduces the amount of virtual machine book-keeping data that needs to be kept in the main memory, making more memory available for desktop VRAM and improving VDI.

The full functions, workings and improvements of VMware View and vSphere can be learned through professional training course such as those offered at Total Training Solutions. Ranging from beginner to more advanced best practices modules, call the TTS team on 0800 612 1299 or email info@tts-uk.com to discuss the course that is right for you.

*Source: VMware View Performance Study Best Practices Technical White Paper.

New Articles

Labels

Keep in Touch