Delivering desktop instances and virtualized applications to remote endpoints relies on fast, secure networking....
By submitting your personal information, you agree that TechTarget and its partners may contact you regarding relevant content, products and special offers.
Network latency and chatty network protocols such as TCP can slow the performance of time-sensitive traffic such as mouse input, touchscreen gestures or streaming content. Horizon 6 integrates a variety of optimizations to improve the performance of common desktop and application content exchanged across the Internet. Collectively, this suite of optimization technologies is called VMware Blast.
From an endpoint support perspective, Blast Adaptive UX increases the communication performance of remote desktops and applications using HTML and a proprietary protocol called PCoIP. Blast Unity Touch improves the way that virtualized desktops and applications run on mobile devices, and Blast Local Access supports access to local USB devices and other peripherals even while users work with virtualized desktops and applications. In the area of media performance, Blast Multimedia improves multimedia streaming tasks such as streaming audio or video, while Blast Live Communications optimizes the performance of real-time audio and video such as Lync or Skype, and Blast 3D handles virtualized 3-D graphics rendering.
VMware can use three technologies to accelerate graphics on servers: virtual shared graphics acceleration, or vSGA; virtual direct graphics acceleration (vDGA); and virtual graphics processing unit, or vGPU. With vDGA, the hypervisor allows VMs to communicate directly with graphics hardware on the server using drivers that are native to the graphics adapter, such as NVIDIA or ATI drivers, for better graphics performance. With vDGA, the VM talks to the native graphics driver, and the VM's workload can perform any graphics function or feature that the graphics hardware can support, including DirectX 11; OpenGL; and even the compute unified device architecture, or CUDA, for parallel computing.
The problem with vDGA is that it's not a shared architecture -- it's only one VM (or desktop instance) per graphics adapter -- so this severely limits the number of graphics-intensive instances that the server can practically support. Even high-end graphics adapters with two or four GPUs pose serious limitations because there aren't enough PCIe slots -- or even power supply capacity -- to operate multiple graphics cards.
Still, vDGA might be the perfect answer when just a few top-quality graphics-intensive desktop instances are needed in the enterprise.
Dig Deeper on Working with VMware View
Related Q&A from Stephen J. Bigelow
Consider factors like security, platform compatibility, data usage requirements and management when transitioning from a private cloud to a hybrid ...continue reading
Several tools and commands can come in handy to storage admins looking to benchmark I/O performance on Linux systems. But not all benchmarking tools ...continue reading
If you consider security, performance, scalability, expertise, network visibility and service management when creating a private cloud, you avoid ...continue reading
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.