Acunu

Acunu is too young to be definitive about its potential success in a crowded market. Much will depend on how successful Apache Cassandra itself is, as well as how Acunu can stand up against other Cassandra distributions. The fact that the company has already displaced a rival at one of its customer sites (the major bank) is encouraging and the tight integration between its analytics and the Cassandra engine gives Acunu a competitive advantage that other companies cannot match.

Unfortunately, we cannot name Acunu's two most well-known customers, one of which is a major bank and the other a major telecommunications provider. Its other users, such as Hailo and Tellybug are mostly in the big data start-up category though Render is an interesting customer, using Acunu to monitor its data centre.

Acunu Analytics runs directly on top of the open source Apache Cassandra NoSQL database. Early adopters of Acunu Analytics typically used an Acunu-supplied distribution of Cassandra because it was more robust and out-performed the standard version of Cassandra. However, the latest version of Cassandra (1.2) is more competitive in this respect, not least because of Acunu's contribution to this open source project. For example, the virtual nodes implemented in this release were contributed by Acunu. Acunu now only supports the Apache Cassandra version for new customers.

Acunu Analytics, as its name implies, is an analytics solutions for building real-time monitoring and analytic applications to run on top of Cassandra. Thus the product effectively combines a database with an analytic development environment. The way that it works is that it incrementally maintains roll-up cubes on data as it is ingested. It reflects fresh data immediately in queries on the cubes, queries that are close to instantaneous because they typically only retrieve pre-calculated results.

Acunu Analytics provides a set of aggregation functions plus associated average and variance operations; count distinct and top-k functions (with variable levels of accuracy) to show things like unique visitors, trending topics, slowest nodes or most valuable customers; and there is support for GROUP BY operators to divide events across the values in a field. It also supports automatic bucketing for time fields and hierarchical grouping to support URLs, paths and geo-spatial queries.

As noted, Acunu offers support contracts for Cassandra implementations even if these were not sold or implemented by Acunu in the first place. For its own product it offers three different levels of support contract, ranging from free best-efforts support only (non-production use) through 8x5, 4 hour response to 24x7, 1 hour response contracts.