This week the FlashSoft team is in Orlando for the Gartner IT Infrastructure and Operations Management Summit and we’re really excited to announce the availability of a new major release of our FlashSoft software for both the Windows Server and Linux operating systems. It’s a major step forward for our software.
Our customers have been using FlashSoft software on Windows Server and Linux for a number of years, to boost application performance, particularly with database applications whose performance is often constrained by I/O latency. In working closely with customers, we’ve learned a lot about their requirements for large-scale data center operations, and in a nutshell, this new product release is all about scalability, flexibility and data protection.
First: Scalability. The caching engine in FlashSoft 3.2 is significantly upgraded. The maximum cache size has been increased from 1TB to 2TB, and since the hot data stored in the cache typically comprises only a fraction of the total data set, this means that a single cache can support data sets of 10TB or more, depending on the profile of the workload.
But there’s more. While previous versions supported acceleration for up to 255 volumes, customers told us that they wanted to use FlashSoft to accelerate applications accessing several times that number of volumes. So we’ve designed the FlashSoft 3.2 software to establish a server-based cache that can concurrently accelerate data access across 2048 storage volumes.
And there’s one more improvement in this release that dramatically increases the amount of data that FlashSoft can support. While the previous version supported a single cache on a server, The FlashSoft 3.2 software can support up to four separate caches on a single server, each on its own solid-state device. So that means that a single server can cache up to 8TB of data on four solid-state devices.
And that brings us to flexibility. With multiple caches, it’s possible to use different solid-state devices for the different caches, to provide the appropriate quality of service to various applications and workloads. With a new LVM built into the FlashSoft caching software it’s also possible to concatenate multiple SSDs in a single logical volume. While this was previously possible with software or hardware RAID, the new FlashSoft-based LVM makes it simple, and eliminates the risk of the RAID introducing latency. By concatenating SSDs, IT managers can incrementally increase the size of the logical volume supporting the cache.
Finally: data protection. The LVM built into the FlashSoft software enables a new capability that our customers are very excited about: the ability to mirror two SSDs to support “safe write-back” caching. For many databases and applications, FlashSoft’s ability to support write-back caching is essential to improving performance. At the same time, in write-back mode data will reside on the SSD for a period of time before being written to storage, so the failure of an SSD can introduce the risk of lost data. By mirroring two SSDs, FlashSoft writes data concurrently to both SSDs, so the failure of a single SSD won’t result in lost data. Most importantly, this mirroring introduces no latency – FlashSoft can write data to the two mirrored SSDs just as fast as writing to a single SSD.
So you can see why we’re excited to be introducing FlashSoft 3.2 this week at the Gartner IT Infrastructure and Operations Management Summit in Orlando. We developed the enhancements in this new release in direct response to our customers’ needs, and we’re dedicated to delivering server-side solid-state caching that meets their expectations for performance, cost-efficiency and flexibility.
To read the full press release, click here.