Actian knows a thing or two about big data, and for over thirty years has provided database management systems that consistently outperform most of the competitors. And so when big data became fashionable Actian already had a head start, and has become the bedrock of big data analytics in many large corporations. While most organizations struggle with Hadoop plumbing, Actian provides the technology and infrastructure to make the big data adventure less hazardous. The speed of its database technology, fully embedded in the Hadoop ecosystem, is a game changer. Businesses can re-optimize their predictive models as frequently as they wish (every ten minutes in some cases), and business users wishing to explore their data can do so with sub-second responses. One organization performs ad-hoc queries on 200 billion rows of data in under two seconds. It makes the dream of any-analytics at any-time a reality, and is something that is quite unique.
Business Application
The essence of the offerings from Actian are speed and flexibility. This enables a host of business applications that would otherwise be infeasible. The iterative processes of building predictive models and data exploration can cycle very rapidly, and both business users and data scientists can gain insights into their data in a way that is otherwise impossible.
Building predictive models is a labor and compute intensive process. Parameters have to be tuned and algorithms selected to arrive at an optimized, reliable model. When the cycle time is measured in hours or days, analyst productivity is severely degraded, costing the organization money and lost opportunity. Cycle time measured in minutes or seconds changes the game entirely, and predictive models can be deployed in a timely manner and re-optimized on a regular basis.
The dream of many business users is to use their data discovery and visualization tools with an absolute minimum of lag between query formulation and display of results. Alas this dream is usually not realized, and particularly on very large data sets. Response times are often measured in minutes and sometimes hours – clearly making many data exploration and diagnostic tasks impossible. The nature of Actian’s database technology means that most queries will execute sub-second, allowing business users to refine and develop their analysis almost in real-time. It’s something we all expect, but is usually not delivered.
Technology
Actian provides a number of big data solutions. Vector is a column based DBMS employing vector based processing (columns can be treated as a single vector). It exploits contemporary hardware architectures, including on-chip cache, aggressive compression and parallel processing. It repeatedly wins benchmark tests – often by an order of magnitude, and performs lightning fast queries without the need for OLAP cubes, schemas and other forms of aggregation. Actian Vector in Hadoop provides very high performance SQL functionality in a Hadoop environment. Actian claims (with some justification) to offer the fastest, enterprise grade massively parallel processing (MPP) SQL query engine to run on Hadoop.
Matrix is also a columnar DBMS that features very fast dynamic compilation of queries, in-memory analytics and is extremely scalable. It is capable of handling very complex joins and optimizes computing resources across multiple nodes.
Actian Dataflow deals with data preparation, analytics development and execution on Hadoop. An embedded data flow engine delivers auto-scaling and parallelism, processing data on Hadoop over 10 times faster than MapReduce. It allows analysts to run analytics directly on Hadoop clusters and dramatically reduces data preparation times.
Actian DataConnect supports data integrations via a visual interface. Many disparate formats are supported and regulatory compliance can be automated.
Predictive model building can be accomplished using KNIME, a platform that Actian has integrated into its product suite. R scripts can also be executed.
About Actian
Actian is a privately owned business that has pioneered high performance database solutions for over 30 years. It is headquartered in Redwood City CA and has offices around the world.