Virtual Desktops That are Cheaper and Faster than a PC.
Importance of VDI Performance
The success of VDI projects begin with the end user experience. Users transitioning from a PC to a virtual desktop expect equal or better performance, yet most VDI deployments provide a fraction of the storage performance required per desktop (measured in IOPS and latency). The most critical aspects of user experience such as boot, logon, application launch and search require vast amounts of storage performance. Often, to lower the cost of VDI, critical features such as desktop and email search are disabled.
|VDI With Atlantis USX|
|Stateless VDI in RAM||Hyper-converged VDI||VDI with Optimized Shared Storage|
|Lower cost than physical PC that outperforms a MacBook Air.
||Simple, scalable and high-performance architecture for any type of VDI.||Accelerates SAN/NAS storage to provide ten times storage performance.|
|Uses RAM as primary storage for virtual desktop images.||Integrated compute and storage solution that uses low-cost local server storage.||Deploy up to ten times more virtual desktops on all-flash arrays.|
Deliver Better-Than-PC Performance—6x that of MacBook Pro
Atlantis USX creates the fastest VDI environment independent of the underlying storage. Atlantis HyperDup™ increases the disk performance of the VDI desktops to 6x that of a MacBook Air with solid state drives. In addition, Atlantis partners have helped to create a library of reference architectures demonstrating the faster than PC performance with lower than PC costs. These guidelines have been followed to deploy very large production VDI environments with over 100,000 desktops.
VDI Costs: Drop your persistent VDI costs to under $300 per desktop
The number one challenge with VDI — Storage Costs
The largest challenge associated with any VDI project is a lack of planning resulting in higher-than-expected storage costs. These extra costs are the results of matching end user performance requirements against the available storage architecture.
In many cases, with SAN/NAS, enterprises have over-provisioned storage capacity in order to achieve adequate storage performance. With all-flash arrays, storage costs are further elevated due to the limited capacity and high cost per gigabyte.
Cost vs Performance
One common solution to reducing costs is to reduce the performance requirements of the system. The throughput of the storage is measured in IOPS with typical PC performance at 80-120 IOPS. Sizing at lower than PC performance means that end users will have a degraded virtual experience compared to the physical experience they are used to. It is not uncommon to see VDI sizing at less than 30 IOPS in order to keep costs down. When the project is scaled out, this creates user experience problems such as slow boot / logon times and sluggish application launches. When comparing costs, user experience must always be a factor of the comparison.
Cut VDI CAPEX Costs
The patented Atlantis HyperDup Content-Aware Data Services eliminates up to 95% of the IO traffic before it reaches the underlying storage hardware, increasing the number of users per storage array by up to twenty times. Atlantis USX operates on any storage, any hypervisor, and with any broker providing enterprise class storage with the fastest performance at the lowest costs.
Without Atlantis USX, costs of delivering VDI with acceptable performance at scale can bring the cost per desktop to $1,000-$2,000. Using Atlantis USX, these costs are reduced by as much as 80% bringing the costs down to less than $300 for persistent desktops and $200 for stateless desktops.
Cut VDI OPEX Costs
The true cost of storage isn’t isolated to the initial purchase, as the yearly operational costs average 40%-60% of the initial hardware purchase. These costs include power, cooling, network, rack space, drive failures, and maintenance licensing. Over the normal 3-5 year of a VDI project, the storage OPEX will be double the CAPEX costs. By using Atlantis USX, the amount of storage that must be maintained and operated is reduced by up to 95%, thereby dramatically reducing the OPEX costs.