It may not be immediately obvious, but today's confirmation of the Oracle/BEA acquisition and announcement of the Sun/MySQL merger do affect the BI market and have significant implications for BI vendors and BI professionals. One look at all of the components needed to create robust, scalable, effective and efficient, "industrial strength" BI solution, and it becomes painfully clear that just about every segment of enterprise software and applications significantly affects BI segment. Sun/MySQL implications are clear, Oracle/BEA are less so, but let's explore them:
Sun / MySQL
- Even though we do not hear of many large enterprise grade data warehouses run on MySQL, Forrester's survey showed that the use of MySQL in enterprises increased from 17% in 2006 to 29% in 2007 (number of enterprises using MySQL in production). It is only logical to assume that Sun will be a significant contributor to MySQL features to enhance its enterprise scalability and robustness.
- This transaction is also another confirmation of the continued success, small but steady, of Open Source BI products such as Jaspersoft, Pentaho, Ingress. Another usage of Open Source model, successfully demonstrated by a major BI vendor, Actuate, is to use open source components as a viral marketing strategy, followed up with selling a commercialized software license.
- After this transaction, watch for increasing number of BI vendors announce tighter integration with MySQL, and watch for Oracle to step up marketing of its own established open-source DBMS offerings.
- Indeed, expect Oracle to tout the fact that it is the leading provider of DBMSs for the open source market. In February 2006, Oracle acquired Sleepycat Software, adding that firm's Berkeley DB embedded open-source DBMS (the most widely used open source database in the world) to its product portfolio. Then in September 2006, Oracle announced the general availability of Oracle Berkeley DB Release 4.5. The product has been enhanced to support multi-version DBMS concurrency and change control; accelerated replication of DBMS instances during development; and the ability to upgrade the DBMS without taking down the entire system.
- This acquisition makes great strategic sense for Sun. The vendor has significantly ramped up its presence in the data warehousing (DW) appliance market recently, mostly through partnerships with vendors such as Greenplum and ParAccel. Most DW appliances run on open-source DBMSs—for example, Greenplum incorporates a massively parallel distribution of the open-source PostgreSQL DBMS. By acquiring its own open-source DBMS vendor, Sun is clearly positioning itself to potentially roll out Sun-branded DW appliances that incorporate and are optimized for MySQL. In its strategic repositioning of its software offerings to the open-source model, it was just a matter of time before Sun made a play for a well-established database vendor. Now, we wouldn't be surprised to see Sun acquire any of the open-source BI players named above, plus perhaps an open-source ETL vendor such as Talend.
Oracle/BEA implications for BI are less clear, but potentially much more significant.
- For the last few years vendors have talked about "pervasive" BI — moving BI from the realm of executives and power analysts to the masses. One approach to making BI more pervasive is to enable universal BI access via Office applications, desktop widgets, mobile. However, easier access to inherently complex applications is only a partial answer.
- BI has traditionally great at answering questions like "what happened", "why it happened" and maybe even "what is going to happen", but it fell short of the next step which is helping us decide "what do I do about it now" and "how do I do it". The latter is the realm of process and rules applications such as BPM (Business Process Management) and BRE (Business Rules Engines). Process-centric BI, or process embedded BI, where dashboards and reports are no longer separate applications, but rather components of a larger process/workflow, has a very significant advantage — it is context aware. In other words, if a dashboard pops up at the right place and time in the process, it does not come up blank, it comes up prepopulated with metrics and attributes, since it knows where it was in the process and what questions you are probably going to ask it. This can potentially solve one of the biggest BI challenges: enabling casual BI users navigate complex BI environments.
- Once Oracle resolves significant BEA product overlaps (which, together with the recent Hyperion acquisition may be a serious distraction) it will end up with a very strong BPM / BRE platform putting it front and center of the process centric BI race. The following products, either overlapping or complementary, will contribute to Oracle's strength in this BI segment: business process management — Oracle BPEL Process Manager and BEA AquaLogic BPM will need to converge, business process modeling and business rules will come from the Oracle side in the form of its BPA Suite and Business Rules
- Other obvious benefits for the enterprise IT users include Oracle and BEA complementary strengths and emphases that promise considerable synergies for their BI customers. Oracle has solution depth and breadth in packaged applications, database management systems (DBMS), data warehousing (DW), extract transform load (ETL), and business performance solutions (BPS). BEA, in turn, has considerable depth in emerging areas such as complex event processing (CEP), enterprise information integration (EII), data federation, and the Web 2.0 "mashup" style of composite application development.
Bottom line: 2008 will continue to be a very exciting, rarely boring, year for BI. No matter how much the market consolidates (I am looking for more BI-related transactions from HP, Sun, Teradata, EMC, Actuate, SAS, Microstrategy and Information Builders) the market is FAR from getting commoditized. I still count over 300 vendors directly or tangentally related to BI, there's not a single week that a new startup with an exciting product comes across my desk, and I can count at least a dozen areas in BI that are still unexplored and unaddressed by large vendors.