Tarmin touts active archiving software
It's ILM, Jim, but not as we know it
Storage Expo UK-based software developer Tarmin Technologies has come out of stealth this week with an intriguing archiving tool called GridBank. As the name implies, this is grid-based and looks rather like ILM version 2.0 - but Tarmin CEO and co-founder Shahbaz Ali prefers to call it active archiving.
Currently in beta and due for full release in January, GridBank runs on a clustered and self-healing grid of heterogeneous servers and has its own global clustered file system, Ali said. He added that it supports just about any storage from any manufacturer, including disk, tape and optical, with the ability to replicate archives from one geographical location to another.
He claimed that ILM didn't succeed because its developers focused on the primary storage, with the aim of pushing data off to other tiers of storage as it ages. "Our focus is on the secondary storage, pulling data off primary," he said. "From a policy perspective it is a very different thing.
"ILM is a process, not a product, and we are continuing some of the processes. The idea is to simplify them and keep as much as possible behind the scenes."
GridBank is structured as two grids, a service-oriented grid of servers which co-operate to share work out, and an object-based data grid built on pools of virtualised storage. Ali noted that its features include content and user authentication, role-based access, policy templates, object encryption, e-discovery, and data-shredding.
The technology had its genesis at Mastercard, where Ali met co-founder Steve Simpson, who is now Tarmin's engineering veep. The pair came up with the idea of overlaying a metabase - a database of metadata - on top of a flat database to overcome the difficulties the financial firm had getting a million transactions an hour into Oracle.
"When I was an end-user, I realised the market wasn't giving me what I needed - a metabase, unified policy management, unified retention schedules, and so on," Ali said.
"Those were the pre-content-aware storage times, then we came across EMC Centera and realised it was something we'd already done - but Centera had no encryption and it was API-based and locked-in," he added.
GridBank uses standard file access techniques - CIFS and NFS - to get at data, then moves and de-duplicates it according to policies, leaving shortcuts behind on the primary storage. ®
Sponsored: Hyper-scale data management