"Decoupling is about moving storage intelligence into servers to get scale-out performance, efficient resource management, data centre design flexibility, and a future proofed storage architecture," said Jeff Byrne, Senior Analyst & Consultant at The Taneja Group. “PernixData has made decoupled storage a reality by creating a clustered software platform that turns high speed server media into a perfect complement to any shared storage environment.”
A New Era of Server-Side Storage Intelligence
PernixData FVP software puts storage intelligence into high speed server media, establishing a control point for optimising application performance and managing data centre operations with true scale-out growth. With PernixData FVP software, customers can leverage any storage platform for capacity, creating a decoupled storage architecture that maximises design flexibility while minimising storage costs.
Unique features of FVP software include hypervisor clustering, topology aware fault tolerance, read and write acceleration, and seamless support for any shared block and file storage systems. In just a short period of time, these capabilities have enabled hundreds of companies across the globe to accelerate approximately 200,000 virtual workloads using a decoupled storage architecture.
The latest version of PernixData FVP software (version 2.5) makes it even easier to deploy a decoupled storage architecture by adding the following new features:
· Memory performance at the price of flash with DFTM-Z adaptive compression. A decoupled storage architecture uses the latest in high speed server media for the fastest VM performance and cost effective scale-out. FVP is the only software that leverages both flash and Distributed Fault Tolerant Memory (DFTM) for this purpose, letting companies deploy storage acceleration on virtually any host.
Given that memory (i.e. RAM) typically offers less capacity than flash, it historically has been used to accelerate only small footprint, low latency workloads like virtual desktops. To make memory cost effective for all applications, including data intensive tier 1 apps like virtualised databases, PernixData introduces DFTM-Z. This is a new feature that compresses data in a DFTM environment, maximising the capacity of the underlying RAM hardware by 4x or more on average. With DFTM-Z, customers can now obtain the performance of memory at the price of flash.
DFTM-Z compression is adaptive based on traffic characteristics, dynamic, and transparent to the workload. This ensures compression overhead never adversely impacts VM performance.
· Optimise what you can, bypass what you must with intelligent I/O profiling. FVP has always had the ability to accelerate on a per VM basis, applying unique acceleration policies to individual workloads. PernixData has taken this one step further by adding a new intelligent I/O profiling function to FVP software.
With this new feature, FVP dynamically detects in real-time which workloads are suitable for server side acceleration and which are not (e.g. large sequential reads and writes). In the latter environment, reads and writes are automatically sent right to shared storage, bypassing local server media. This not only maximises the performance of these workloads, but it maximises the performance of workloads using server side acceleration by ensuring server flash and RAM is available for storage acceleration when needed.
· Granular visibility with Role Based Access Control (RBAC). FVP software is increasingly being installed in Private Cloud deployments and from Service Providers with Infrastructure as a Service (IaaS) offerings. To facilitate this trend, PernixData recently introduced a subscription version of FVP software to align with managed service offerings. In addition, these environments usually have multiple tenants who desire unique views into their own infrastructure and application environments. To support this functionality, PernixData has added Role Based Access Control (RBAC) to the latest version of FVP software, letting administers control who has access to the PernixData User Interface (UI) and what information they can see.
In addition, RBAC can be used within an IT department to give granular access to acceleration data. For example, a Database Administrator (DBA) can be given visibility into database statistics while excluding visibility other VMs.