The hard truth in end user computing is that the upfront costs of a VDI deployment are often higher than purchasing physical workstations. By the time you tally up the new data center gear and software licenses you’ll need, you might have spent less upfront than buying everyone a new laptop. The real drivers for adoption in this fast growing category are the ongoing benefits – greater operational efficiencies, improved security, and workforce mobility, to name a few. That being said, none of this means cost containment isn’t a top priority.
At SimpliVity, we take a very unique approach to keeping costs down for a superior end-user experience, while still delivering unmatched performance and enterprise-class resilience.
SimpliVity’s unique data architecture allows us to achieve the highest desktop density in the hyperconverged infrastructure category. We’ve shown in rigorous third-party performance testing that we can host up to between 250 and 300 users per OmniStack system in both VMware and Citrix environments. How do we do it? The OmniStack Data Virtualization Platform eliminates redundant I/Os and dramatically reduces the disk capacity required to host the average desktop virtual machine, by performing inline deduplication, compression, and optimization of all data at inception, across all storage tiers. Since we leverage our OmniStack Accelerator Card, all this happens without taxing the production CPUs – a huge plus in production VDI environments.
In working with SimpliVity Solutions Architects to design and implement end user computing solutions in the field, one thing we’ve seen in customer implementations is that they may run out of compute capacity before we run out of the ability to service I/O for additional desktops. In these cases, a great way to keep costs down for the solution is to leverage what we call “compute nodes.” Compute nodes are simply non-hyperconverged servers that plug into the SimpliVity solution and provide additional compute capacity while leveraging the storage provided by the OmniStack infrastructure. Effective use of compute nodes in a properly designed environment can provide best-of-breed performance for delivering desktops and applications to your end users while driving down cost per user.
Another great use case for compute nodes is users with 3D graphics requirements. Consolidation ratios for users who need graphics processing units (GPUs) has traditionally been very low. By using compute nodes, we can get more physical GPUs into the environment, driving up consolidation while we drive down the cost per 3D-supporting desktop. Plus, we’re able to support the latest GPUs, such as the new NVIDIA GRID M10, as soon as they’re qualified by the server vendor.
In the end, our OmniStack Data Virtualization Platform is the secret sauce that makes our hyperconverged solution ideal for VDI deployments. By extending that storage platform for use with compute nodes, we simultaneously drive down cost per user while enabling advanced functionality for our customers.
Want to learn more? Check out some of our latest performance testing with Citrix XenDesktop to see how SimpliVity gets the highest density of desktops among any hyperconverged vendor.
About Jason Shiplett
Jason is an IT architect with 15 years of industry experience, mainly focused on server virtualization and end user computing. He is VMware Certified Design Expert (VCDX) #183, certified in both Datacenter Virtualization (VCDX5-DCV) and Desktop and Mobility (VCDX-DTM). As a principal solutions engineer at SimpliVity, Jason designs and validates VMware EUC solutions on SimpliVity OmniStack infrastructure.