This article is more than 1 year old

Caringo insists its software is more than a Swarm in a tea cup

What's new in version 7.5

Object storage startup Caringo has upgraded its Swarm software to handle partial disk failures gracefully and play nice with large object deletions, node retirement, and upgrade.

Swarm is Caringo's re-branded and developed CAStor object storage software and came out as Swarm v7 in May last year. Caringo said it was going to offer object, file, block, and Hadoop support "out of the box," and no longer be an object storage technology accessed as object storage.

Instead it was moving to be an all-inclusive scale-out storage technology with object storage underpinnings. Swarm HadoopFS, a native connector for Hadoop, was announced in October last year.

Now we have Swarm v7.5 which:

  • Retires and refreshes hardware more efficiently
  • Keeps disks online when partially failed
  • Adds Multipath support

It adds processes to eliminate potential performance slowdowns and saves time when upgrading to new hardware, adding capacity, or retiring failed drives. This limits the impact of capacity balancing during hardware maintenance and refreshes. Caringo says "a single-step synchronous process now makes data swaps more predictable and efficient. Users can limit operations to specific nodes, providing greater control during hardware maintenance operations."

This helps when large objects or a lot of objects are deleted, such as digital video records that have outlived their usefulness, or moved when old hardware is retired from a Swarm cluster.

Swarm 7.5 monitors disk health in such a way as to keep a drive online when it can no longer write data but can read it. The disk can then be replaced with less performance disruption.

Multipathing increases IO capacity by supporting dual SATA connections. Swarm also gets support for "newer, denser storage arrays for high availability, high performance, and seamless migration to web-scale capacities."

Tony Barbagallo joined Caringo as VP for products in June, coming from being chief marketing guy at Clustrix, previously VP of Marketing at WD-acquired Skyera, and a CV including stints at EVault, EMC's Dantz, and Microsoft. We should expect more product news – a canned Barbagallo quote said, "This is the first of many innovative enhancements and products we have coming out in the second half of 2015."

Caringo co-founder Paul Carpentier left his CTO position in August last year to join Austin Ventures, but stayed on Caringo's board.

When object storage was just that, object storage, it was a separate storage tech set apart from the mainstream with its own use cases. Now it is being presented as generalized multi-protocol storage better suited for large, web-scale, generalized storage needs than monolithic or dual-controller arrays.

As such, it must compete with non-object-based, software-defined storage offering a similar inclusive and multi-protocol storage facility, either combined with app processing in hyper-converged systems or presented as a better kind of shared storage for all kinds of secondary data. Swarm is an example of the latter.

We're told existing Caringo customers can find out more about Swarm enhancements in the product doc available in Caringo Connect – login required. ®

More about

More about

More about

TIP US OFF

Send us news


Other stories you might like