We’ve touched on this topic before, and as the march continues to a fully virtualized data center and the movement of applications and data to the cloud it’s becoming more apparent. It’s also related to our roll-out of virtual storage appliances that we announced this week.
Riding a wave of server virtualization technologies, the compute infrastructure is by and large built on a scale-out commodity hardware layer running a variety of applications in virtual machines. Resources are efficiently shared by multiple applications and users in an environment that is centrally managed, automated for large scale, and extremely responsive. Open source software has a prominent place (LAMP stack anyone?). It’s the model pioneered by the web giants when they were the only ones facing challenges of that scale and now it’s ubiquitous.
Storage is increasingly becoming a commoditized, virtualized, centrally managed pool of resources – in short it’s transforming to reflect the compute environment in many respects. We can again look to web giants such as Goolge who has built their own storage infrastructure that has all of the cloud-like characteristics: built on commodity scale-out hardware, the intelligence and functionality are built into the software, it’s very dynamic, scales on demand, and gives high availability through redundancy. The problem is there aren’t many companies that can bring to bear the engineering resources of Google to build something from scratch. This is why at Gluster we are working to accelerate this storage transition and provide enterprise customers Google-like storage off the shelf.
Which takes us back to delivering Gluster as a virtual storage appliance. It borrows lessons from the server industry, takes another big virtualization step, and enables storage to be deployed the same way you deploy virtual servers. Also, it can be deployed in your own data center or in the public cloud. Feel free to give it a try.