OK, so we’ve heard that question a few times. And we’ve been listening. It’s not like we didn’t support Hyper-V before. You could run FlashSoft for Windows on a Hyper-V server and accelerate storage I/O. The application performance gains are actually very good (see the chart below). You just couldn’t do Hyper-V cluster-y things. Like LiveMigration, or HA.
In FlashSoft 3.7, we’re now letting you accelerate Cluster Shared Volumes (CSV). And that gives us Hyper-V cluster support. It obviously wasn’t that simple to do, but it is pretty simple to explain. When you create your virtual machine, you target a CSV. And when you configure FlashSoft, you target that CSV for acceleration. Done! Now all of the virtual machines on that CSV are accelerated. If you migrate one of those virtual machines to another machine, if it has FlashSoft and it’s been configured to accelerate that CSV, then your virtual machine will continue to be accelerated. And if you go back to the original server, FlashSoft knows not to use the (possibly stale) old cache data. So now my data is safe and my applications are performing better. Pretty sweet.
We’ve got some more ideas planned for Hyper-V. We’d like to provide more granular cache selection – maybe at the virtual machine level or even the virtual disk level. And we would like to support NAS environments as well as SAN. But we’re very excited about what you can do now with FlashSoft and Hyper-V in version 3.7. Give it a try and let us know what you think.
What else is in 3.7?
FlashSoft 3.7 is not just about Hyper-V. We’ve also done some clever things in the “bare-metal” space. For example we’ve completed our support for PowerShell in FlashSoft for Windows. Now there is nothing you can do in the GUI that you can’t also do with PowerShell. Scripters and GUI-haters rejoice.
In our Linux version we’ve also been giving a nod to the scripters. We’ve updated our command-line interface significantly and made it much more “standard Linux”. I’m not sure scripters will exactly rejoice, but I’m hoping they curse us less often.
In Windows 3.7 we’ve also made a big QA effort around Windows Failover Clustering. We knew that nothing in FlashSoft (in write-through mode) prevented you from using Failover Clustering. But we decided to do some serious qualification around SQL Server and Failover Clustering.
SQL Server is a key use case for caching as you can imagine. And we’ve also spent some time with SQL Server AlwaysOn Availability Groups. Again, there is nothing in FlashSoft to prevent using it in this way. As our Chief Architect says, “there is no reason it should not work”. But for some reason our customers wanted us to say something a bit stronger. So we are.
We tested FlashSoft with SQL Server AlwaysOn Availability Groups and it works. It works really, really well. Keep an eye out for the forthcoming performance reports and a Solution Guide.
We also wanted to work really, really well with our new corporate partners from the FusionIO acquisition. FusionIO, with their ioMemory3 cards, was one of the leaders in bringing devices formatted with 4k sectors to market. FlashSoft now supports both 4k sectors and 512 sectors in any combination. So you can use a 4k SSD as a cache for a 512 formatted backend storage device. Or 4k and 4k or 512 and… you get the drift.
On the Linux side, we also want to continue to support a very wide range of distributions. We added back support for RHEL 5.5-5.11 due to popular demand. At the same time we added RHEL 7 for the leading edge types. And we are supported SLES 11sp3 and 12.
And as always we’re always tinkering with the UI to make the software easier to install and configure and to figure out if it’s working.
We’ve been busy, and we’re very proud of what we’re releasing now. We hope you like it.