Actian goes on a Big Data acquisition spree
Actian, the vendor formerly known as Ingres, has suddenly put itself on the map through rapid-fire acquisitions of Pervasive Software, ParAccel, and Versant Software. The die was cast more than two years ago with its acquisition of Vectorwise, a fast, single-node analytic database.
Actian’s goal is to assemble a mini-stack that includes a data integration engine and different data platforms for different transactional and analytic use cases. With the acquisitions, Actian winds up with a mix of mature, maintenance-oriented businesses that can contribute cash flow towards developing emerging data platforms for harnessing Big Data.
While Actian has bought its way to a $150m business – big enough to be noticed – its challenge is building critical mass in at least one or two of its product segments.
Actian, the outgrowth of the Ingres business, has been in search of a high-growth engine since the company changed its name in 2011. Capitalizing on the Vectorwise acquisition, its initial focus was cloud-based “Action Apps” that would be triggered through realtime changes in data. While the company still harbors cloud-based aspirations, the current focus is more on diversified data platforms united by an integration technology stack.
The company’s business originated around the venerable Ingres relational database, an engine that originally used a different query language that competed with and ultimately lost the battle to SQL in the 1980s. (Ingres eventually added SQL support.)
Through a series of acquisitions, Ingres wound up in CA’s portfolio, and in 2005, CA spun it off as an independent company with the technology available in open source and enterprise versions. Since then, Ingres, now Actian, survived on license and maintenance revenue from a sizable legacy installed base.
In 2010, the company bought Vectorwise, an SMP (symmetric multiprocessing) columnar analytic database that competes with Sybase IQ, but with an added twist: it “vectorizes” processing at the chip level, and is capable of applying a single instruction set to multiple data points (also known as “Single Instruction, Multiple Data” or SIMD).
With its high-speed SMP architecture, the sweet spot for Vectorwise is data sets between 1–10TBytes. As such, it is sandwiched in the market between the better-known Sybase IQ (which has been around for nearly 20 years) in the SMP columnar space and, at the other end, by more scalable MPP platforms (e.g., Aster Data, now owned by Teradata; Greenplum, acquired by EMC and now spun off as Pivotal; and Vertica, acquired by HP).
Among recent acquisitions, the ParAccel deal at first glance appears to be redundant; our initial question was why Actian needed yet another columnar analytic data store. Ultimately, the answer was that ParAccel and Vectorwise each provide missing pieces that the other lacks.
Combine Vectorwise’s on-chip, SIMD process capabilities with ParAccel’s scale-out, and you have a lightning-fast platform that could conceivably scale out to petabyte levels. Prior to the acquisition, ParAccel was the last of the first wave of Advanced SQL analytic players that had not been acquired by deep-pocketed IT household names. Amazon used a subset of its engine to power the new Redshift MPP analytic database service now on AWS, but didn’t publicize the ParAccel connection. Despite receiving tens of millions in venture backing (including a sizable investment from Amazon), ParAccel’s market presence remained fairly static.
The hidden gem of the recent acquisitions is Pervasive, which already had an established data integration business that competed with Informatica in the midmarket. The $162m deal held little mystery, having been negotiated since the middle of last year.
Pervasive’s data integration and data cloud (integration services that are available on AWS) potentially form the glue between Actian’s increasingly diverse array of data platforms. Pervasive also had another budding asset that is the key to Actian’s future Big Data strategy: DataRush, a patented framework that automatically multi-threads processing jobs based on the number of cores that are detected at runtime.
Pervasive Software claims that DataRush has higher benchmarks than MapReduce, and promotes it as an alternative, high-performance framework for processing data on Hadoop and as a data pipeline for Actian’s SQL analytic platforms.
Capping off the recent spree of acquisitions was Versant, an age-old, object-oriented database (OODB) provider that has survived on a niche market with telecom clients. Back in the 1990s, OODBs were hailed as the logical successor to SQL databases because they overcame the “impedance mismatch” between multi-tiered software architectures and the data.
But SQL was too entrenched, relegating OODBs as a footnote in database history. But OODBs have resurfaced, not as enterprise data platforms, but as specialized data stores such as the Java-based in-memory data grids that provide workarounds to I/O database bottlenecks for high-traffic, web transaction apps; and the new generation of web developer-friendly document-oriented NoSQL data stores.
Making 1 + 1 = 3 is the challenge
On paper, Actian has achieved critical mass in terms of overall revenues and now is diversified enough to become a one-stop provider of OLTP, operational, and analytic data platforms. But that skips over the reality that Actian’s data platforms are niche offerings, none of which have critical-mass presence in their respective markets, and that enterprises are more likely to rely on Actian to provide alternatives to incumbents such as IBM and Oracle.
Actian is competing in a database market that is newly receptive to upstarts and populated by plenty of competition for Fast Data and Big Data use cases for transaction processing and analytics.
Actian needs focus; it must forge critical-mass presence in one or two key product markets. That dictates converging some of its products, as maintaining a proliferation of niche platforms is simply not viable for a $150m company (product proliferation will stretch the company’s finite product development and go-to-market resources). The most promising opportunities are:
- building off Pervasive, which has an established data integration presence in the midmarket and an emerging one with cloud computing and Hadoop; and
- fusing the ParAccel and Vectorwise technologies to produce one wickedly fast and scalable analytic platform that can complement Hadoop.
- » VMware acquires Bitnami to accelerate multi-cloud and enterprise deployments
- » Unravel Data secures $35 million series C funding to help combine APM with big data muscle
- » Google’s cloudy head count and revenues remain on the up – but specifics are still a while off
- » Exploring WAN data acceleration: Is edge computing really necessary?
- » How to ensure your SMB's cloud adoption is painless: A guide