Birst Inc., an enterprise software company not named after anything that explodes, Dec. 14 launched the first cloud-based data analytics service that deploys an in-memory database specifically tuned for analytics.
The SQL-based, columnar database service housed in DRAM (dynamic random access memory) works fast and handles large workloads, enabling users to turn previously unused enterprise data into intelligent analytics reports that can immediately improve decision-making at many levels of a company's management.
Because it runs in real-time DRAM, and not on hard or solid-state drives, Birst's analytics are unconstrained by the limits of single-table or associative models used by conventional discovery tools, Brad Peters, CEO of the 6-year-old San Francisco-based company, told eWEEK.
Bringing the BI Stack Into the Cloud
"What Birst is doing is bringing the business intelligence stack into the cloud," Peters said. "The previous generation of business intelligence analytics was all well and good, but often had to be cobbled together with various parts. SAP has their stuff, IBM has Cognos, Oracle has Seibel analytics, and so on. They were designed for a previous generation [of IT system]. You had to bring together all the parts to make meaningful decisions based on facts and data.
"The hurdle for building a BI implementation is really high. What we set out to do is build a virtual appliance for BI, such that the barriers for adopting BI are dramatically lower."
The Birst in-memory analytics database uses the company's own home-developed data warehouse automation IT, which provides for data integration across numerous data sources. These include such oft-used data streams as SAP, Salesforce, and from various operational and financial systems and incorporates them into a multidimensional, star schema design.
Because it is based on the universal database Structured Query Language, SQL, and open-source Java, Birst's database is interoperable with existing applications and allows for open access via Java Database Connectivity (JDBC). Columnar structure, compression and parallel processing speed performance for analytics and achieve high throughput, Peters said.
Birst's cloud service also enables analysis of data that is modified in real time, so that it empowers business users to make critical business decisions using the most current data.
"This is a real relational database that allows for complex (e.g. multipass) and ad hoc analysis of sophisticated relationships not possible in simpler tools," Peters said.
Alternative to Oracle, SAP, et al
Birst is an alternative to conventional business analytics, such as SAP's HANA and Oracle Exalytics, Peters said.
The Birst in-memory database is bundled at no extra charge in subscriptions to its cloud analytics service. The company licenses according to numbers of seats and per-gigabyte of random access memory utilized. For example, a 200GB, 10-seat implementation will run a company from between $30,000 and $35,000 for a one-time license, Peters said.
The database will become generally available in January 2012, Peters said.
Birst is staging a live Webinar to demonstrate the in-memory database on Thursday, Jan. 19, at 10 a.m. PST. Register here.