benefit from all our premium research
related research from the MWD library
most recent posts
Wednesday, February 12, 2014 by Helena Schwenk
Actian (formerly known as Ingres) recently held a roundtable event in London moderated by the company’s CEO and CTO, Steve Shine and Mike Hoskins respectively. The gathering was designed as an open discussion between press, analysts and Actian and as a forum for exchanging ideas and feedback about perspectives on Big Data and the role Actian wants to play.
Starting the year with such an event made sense. The company had a busy and re-defining 2013, having recently made several acquisitions, including ParAccel (for its MPP analytical database), Pervasive Software (a data integration engine) and Versant Corporation (a NoSQL object database) – and not forgetting Vectorwise for its SMP analytical columnar database which was acquired earlier in 2010. Apart from doubling the size of the company in the last year, both in terms of people and revenue, these acquisitions were also designed to help round out Actian’s Big Data and analytics platform and position it more competitively with the other specialist Big Data players in this space. Having spent a significant proportion of 2013 integrating technologies and launching its resulting offerings it seemed only wise to kick start 2014 by setting out its Big Data stall.
While the roundtable didn’t necessarily have a set format or structure, a good proportion of the conversation focused on the role Hadoop plays in the Big Data ecosystem. Actian, of course, had a lot to say on the matter. It believes Hadoop and its associated technologies are great for advancing the use of large scale and unstructured data, but (and there is always a but) – up until a point; to put it more bluntly, the company thinks it’s great for scale but not necessarily great for performance. Actian has a point. Hadoop was never designed for real-time processing or interactive querying, two capabilities sought after in the BI and analytics world where business users demand information at their fingertips in increasingly shorter timescales.
That said, most agree that Hadoop does have a very valuable place in today’s data warehousing and analytics environment. It is, for example, very good at offloading the problem of ingesting, processing and delivering data on a massive scale – whether it’s structured or unstructured – at a lower cost per terabyte than traditional analytic databases. It manages this through its massively scalable storage and batch data processing system that scales horizontally on commodity hardware and provide fault tolerance through software.
Like most other analytical database vendors Actian has rather wisely chosen to embrace Hadoop, rather than ignore it; but at the same time it’s looking to address some of these performance shortcomings. Its acquisition of Pervasive and ParAccel in particular are helping to provide a viable alternative to those seeking either a more performant data integration engine on Hadoop (amongst other platforms) and/or want specialist databases suited to different analytic workloads.
With its growing set of Big Data technology assets, a key challenge for Actian going forward concerns its ability to articulate its value proposition and the use cases for each platform. The company still needs to work at establishing a name for itself in the Big Data market, having rebranded from Ingres in mid 2011; but in addition it also needs to work hard at outlining where and how Vectorwise and ParAccel fit into this landscape.
While there might be some cause for confusion over the need for two analytic platforms, both do largely complement each other. Vectorwise leverages the compute power of processor chips to provide a low latency, high performance platform up to the ten terabyte range, whereas ParAccel scales out and provides a higher performance analytic environment on much larger data volumes. If Actian ever pursued the opportunity to combine both in one platform that that would make for a very compelling offering.
The other special ingredient thrown into this mix is Pervasive, which can help provide a much needed bridge between these and other Big Data platforms. Its Dataflow offering, for example, can be used to replace Mapreduce to support a higher-performance framework for processing data on Hadoop where massive scale is required, before the data is potentially piped to other data platforms – helping organisations piece together the Big Data platforms technologies of their choice. In the end though, Actian’s assembled set of Big Data technology capabilities together with its ability to work with Hadoop provide a robust and viable alternative, especially for those mid-sized organisations that might have been previously priced out of the Big Data game.