3dmentat - Fotolia
PernixData FVP boosts RAM as host storage with up to 4x compression
Flash caching firm PernixData adds compression to FVP 2.5 for more scope to use RAM as host-based storage for high-performance datasets
Flash caching software provider PernixData has upgraded its FVP product to include RAM compression and the ability to offload workloads unsuited to solid state acceleration to SAN or NAS shared storage.
The upgrade – which takes FVP to version 2.5 – will give customers up to a 4x capacity boost when using RAM as a cache, said PernixData marketing vice-president Jeff Aaron, and can be used for all types of workload.
PernixData FVP is one of a number of flash caching software products. It runs on the VMware hypervisor, intercepts database writes and routes them to local host-based flash storage or RAM. That hot data is then available to applications, with extremely rapid access, while data is also written to the storage array.
Flash caching products have arisen alongside the rise of flash storage. As customers put flash storage into servers to boost access times to hot data, software products have emerged that can help route that data efficiently.
Read more about flash caching
- Definition: Flash cache
- Tiering vs. caching in flash-based storage systems
- Server-side flash cache pros outweigh cons for many
- Write-through, write-around, write-back: Cache explained
- Flash caching software market roundup
RAM compression widens options
The new feature on PernixData FVP – called DFTM-Z – provides compression to RAM. It allows customers more scope to use on-board memory as cache where previously they may have shied away from it, because there wasn’t sufficient capacity on host RAM to hold important datasets. It gives customers the option to use RAM instead of flash, which also provides low latency but not as rapid access as motherboard-mounted RAM.
PernixData has also added Intelligent I/O Profiling. This identifies workloads and sends those unsuited to caching directly to shared storage SAN or NAS arrays. This allows PernixData caching to be directed only at the datasets that can benefit from I/O acceleration.
FVP creates a form of write-back cache, where data is acknowledged to the host before it is written to permanent storage. This can risk data loss in cases where a server fails but, with PernixData FVP, customers can specify that data is written to more than one cache to mitigate against host failure.
