With today’s modern large computer, network & storage infrastructures, rarely do we talk just about single product, or even a single vendor’s solution being able to solve the many business challenges being faced. One important area that the Citrix Converged Infrastructure team puts considerable engineering effort into is the hypervisor layer, particularly around its features, performance and scalability.
Focusing for a moment on application and desktop virtualization use cases from the [Citrix XenServer] hypervisor perspective, there is a great deal we can do to drive increased value for customers’ infrastructure projects.
The Converged Infrastructure group has been very busy this year making great strides in this with many new features and increases in performance & scalability in the January v6.5 release and more recently with the May SP1 release, which includes additional new features.
Additionally, NVIDIA GPU pass-through for Linux and Intel GVT-d, GPU pass-through for Windows are now also supported configurations. The Intel GVT-d support is particularly interesting as there is no extra hardware required with the embedded GPUs enabling enhanced graphics for virtualized applications & desktops.
With the greater increase in scale and performance–and of course the enhanced graphics capabilities with an ever-increasing number of consumers–there comes the need for IT admins to be more aware of their infrastructures and of how all aspects of the infrastructure from computer to storage and networking to graphics are performing together and inter-dependently.
Citrix XenServer offers some of this vital monitoring and reporting, either from the XenCenter performance tab (at host & VM levels) or via the Workload Balancing appliance feature; however, for larger and more complex infrastructures, customers may want a single performance monitoring and reporting solution that encompasses not just the hypervisor layer, but also other areas of the infrastructure, such as applications and desktops (XenApp & XenDesktop) as well as application delivery controllers & load balancers (NetScaler).
Released on 24th June 2015, it leverages XenServer v6.5 to bring real-time visibility into NVIDIA GRID, storage performance and XenServer resource provisioning in one console.
Take a customer scenario, where an IT admin is monitoring the infrastructure relating to an application or desktop virtualization performance problem and where HDX/HDX 3D is in use. Visibility across a number of products, including being able to track what GRID card’s CPU and memory resources are being utilized, and then being able to drill down to the utilization at the VPU level for each session to see which VM or VDI is consuming all of the resources is incredibly useful.
With XenServer, it is now possible to correlate the GRID card performance with a particular VM or VDI and understand vGPU utilization and an overall user’s session performance across multiple layers of infrastructure.
Key proactive and real-time troubleshooting features include:
- NVIDIA GRID GPU & vGPU performance trending, alerting, and reporting including CPU, Memory, Power, & Temperature
- Storage performance trending, alerting and dashboards: IOPs, Latency, I/O Wait, and Throughput
- XenServer Host and VM CPU, CPU Load, Memory, Network, and Latency
- Citrix XenApp and XenDesk end user experience per session and application
Citrix application and desktop virtualization solutions scale to meet even the largest customer infrastructure and performance demands. With Citrix XenServer and Goliath Performance Monitoring, customers can proactively monitor the performance of applicable infrastructure layers, now including their NVIDIA GRID sessions.