Blog Post

Exclusive: Big Data Startup Acunu Raises Small Funding

Data storage startup Acunu has closed on a £2.2 million ($3.6 million) Series A round to help bring its first product to market. The London-based company raised the money from Eden Ventures, Oxford Technology Management and Pentech Ventures LLP, and the round brings Acunu’s total funding to £3.1 million.

Acunu makes software that will help solid state drives act as a more efficient bridge between expensive, in-memory RAM and cheap-but-slow hard disks. Instead of writing to hard drives and leaving information locked away, today’s applications want access — and quickly — to a much larger percentage of data stored in memory or on memcached servers or SSDs near the web server. But that speed comes at a cost, and the closer to the compute the storage is, the more expensive it is. SSDs are emerging as a good compromise, but Acunu believes there’s a niche it can fulfill by rewriting the software stack for SSDs.

The Acunu software runs on commodity servers, and appears optimized for solid state drives. It’s in private beta with a few undisclosed companies in the social media, advertising and cloud computing sectors participating. The software kernel uses a modified version of the Cassandra database for input and allows for multiple queries on distributed databases, giving one a unified view of the databases, which is key for companies with multiple data stores around the world.

For more on the topic of big data, including some hot startups, I hope to see you at our Structure: Big Data event on March 23 in New York.

Related Research about on Big Data from GigaOM Pro:

3 Responses to “Exclusive: Big Data Startup Acunu Raises Small Funding”

  1. Acunu is doing somewhat more than a flash cache; it’s rewriting the storage stack for Big Data. This is a fundamental bottom-up rewrite, including advanced algorithms for RAID, LVM, buffering, caching, prefetching, and of course the FS – which we believe is the wrong abstraction for Big Data applications. Instead of exposing a file system, it exposes a sort-of generalized key-value interface so that applications such as Apache Cassandra and others finally have a storage stack on which to build. No more userspace FSes – this is the storage stack for Big Data.

  2. The “more efficient bridge” between RAM and HDD is SSD cache:

    Solaris ZFS L2ARC or Facebook’s flash cache for Linux. It would be nice to see something more interesting of course than these obvious caching solutions.