VDI makes sense, at least from IT’s perspective. It can provide some cost savings in hardware and software, can improve operational efficiency and deliver better control over the company’s data. For these reasons the IT manager and the CIO love VDI; but users often feel very differently. Replacing a desktop computer with a virtual desktop doesn’t always thrill the employees. And if they don’t like their virtual desktops, the VDI project is doomed to failure.
Want your users to love their VDI environment while you still save budget and improve efficiency? Watch our webinar, available on demand, to find out how.
VDI – What’s not to like?
As we mentioned above, it’s not hard to see why IT and management are big fans of VDI, but too often employees don’t share that sentiment. The reasons users most often cite for this dissatisfaction is poor performance and a lack of personalization. Applications run too slowly, especially at peak times, or they don’t look and feel like their own personal desktops.
The reason for poor performance is typically an inadequate storage infrastructure, especially when it comes to sustaining high write transactions. On personalization, users want the same look and feel with their virtual desktops that they had with their physical computers. This means implementing persistent desktops, something many companies choose not to do, in order to reduce cost and keep performance requirements in line.
How to keep users happy
On the performance issue, the VDI infrastructure needs to be supported by a storage system with enough IOPS to handle the hour-by-hour workload generated by the VDI servers. But is also must handle the peak times, like the beginning of the day, the end of the day, when applying updates, etc.
This doesn’t mean just throwing a bunch of flash at the problem. While solid-state devices (SSDs) are typically part of the solution, there’s more to it than that. How does the system handle overhead like thin-provisioning, deduplication, cloning and other processes that tax performance? These functions can increase effective capacity and are often key to delivering the economics of a VDI system, but consuming performance to get them is not a good trade off.
Performance and Persistence
There are new technologies available from companies like Atlantis Computing that can improve ‘capacity efficiency’ without affecting ‘performance efficiency’. They do this by optimizing VDI read and write traffic before it gets to the flash tier, typically leveraging DRAM that’s available in most servers to augment that cache capacity.
Persistence means using persistent desktops. This can increase storage consumption if each user’s desktop is allocated its own capacity, which can add up when you’re talking about hundreds of users. Thin provisioning and clones can help address this capacity problem but may create a performance issue in return, since those features further increase the demand for write performance.
As mentioned above, using software that can optimize read and write activity and coalesce the data stream before it’s written to flash can help maximize performance. It can also leverage ‘content awareness’ to make sure the right data is in cache at the right time.
To learn more about why users often dislike their VDI desktops and what to do about it, join Storage Switzerland for this on demand webinar. In it we discuss how the right storage solution can make your users love VDI.