PernixData FVP 1.0 is arrived

0 Flares Twitter 0 Facebook 0 LinkedIn 0 Email -- 0 Flares ×

Since its first public appearance at the Storage Field Day 3 event I attended on last April, where CTO Satyam Vaghani told us about his company, their technology, and demoed us their produtct, there has been a huge hype around this young startup.

PernixData has only one product, called FVP (Flash Virtualization Platform) that allows to join together in a “cluster” the several SSD or Flash cards installed locally on every server in a vSphere cluster, creating what could be called a “Cache Area Network”. This new storage layer accelerates both reads and writes from the hypervisor to the backend storage, speeding up greatly the overall performances:

How PernixData FVP works

The basic idea is to completely separate Storage Capacity from Performances, with the latter totally managed by FVP. In this way, you can now design your backend storage only by their needed capacity. Even “slow” storage systems can now reach performance levels originally out of their reach, by simply using FVP inside the vSphere cluster. This technology is disruptive in the design practices you can now apply: vSphere architects can now design around a multi-layer storage totally different from the past.

With the announcement of the first GA version 1.0, you can finally test yourself this product by simply request a 60 days trial here, and we also finally have some informations about the price. While I was partecipating into the beta program, I tested FVP and I was amazed by the performances and also how it perfectly integrates inside vSphere.

About the price, the official value is 7500 USD per licensed host, regardless cores or VMs each hypervisor has. There will be also in the future special packages for SMB and Service Providers. If you think about how much a storage vendor charges you for an SSD disk to be installed inside their arrays, and also about the fact this solution does not offer the same level of performances of an SSD directly installed inside the hypervisor, you can realize how PernixData can really be a great alternative for designing storage, or even a way to extend the life of a SAN that maybe has no more performance left but it’s still plenty of disk space.

During the next VMworld US in San Francisco from 26th to 29th of August I will try to meet again Satyam and the other guys at PernixData to have feedbacks about the new 1.0 version and maybe some information about the upcoming roadmap.