(Why would I end a meal with nuts? That phrase never made sense to me.)

IBM made some announcements on April 28 related to the Big Data space. Amongst those announcements were:

  1. The addition of Big Data/Analytics Services onto the Bluemix development platform – time-series dB, embedded reporting (built upon its Cognos offering), geospatial analytics, Internet of Things (IoT) APIs for data access and collection).
  2. A newer version of InfoSphere InfoStreams for enhancing its data-in-motion analytics, specifically with IoT.
  3. BlueInsight: its “vision” of Big Data related services to be available on Bluemix, enabling enterprises to build out analytics apps (to be released at the end of June 2014).
  4. BigSQL 3.0, an alternative to the MapReduce approach, allowing users to run SQL-like queries on Hadoop as done with RDBMS (also to be released at the end of June).
  5. A new line of servers based upon the recently announced POWER8 processor. IBM stated that POWER8 was specifically designed for Big Data workloads, while enabling consolidation of Big Data scale-out architectures.

My reactions to these announcements?

  • IBM is stressing “user friendliness”, perhaps to encourage adoption of its Big Data offerings. Its BlueInsights vision wants to enable the “data-centric” professional to build its own analytic apps (using services based upon the varied data analysis tools already in its portfolio) – a PaaS for Big Data Analytics.
  • The BigSQL 3.0 announcement also signals “user-friendliness”, as this version will allow the results of SQL queries on both database types can be pooled into one dataset for analysis.
  • IBM wants to grab early mindshare in the Internet of Things space. On April 23, Google showed off a server motherboard designed with POWER8 chips. This does indicate that Google is considering a shift away from x86 servers that dominate datacenters today. Perhaps Google anticipates the next huge wave of data coming from its foray into IoT (Nest SV acquisition, Google Glass) and wants more powerful servers to address the anticipated processing needs. And perhaps IBM wants to take advantage of this and position itself as a solution provider for IoT.

And yet, I still have many questions. (Duh, I AM an analyst…)

First, will the BlueInsights component of BlueMix offer enough functionality for the big data professionals to create meaningful and usable analytics apps? Will IBM’s target customers want to make their own apps or go for an off-the-shelf solution?

Second, how early is too early for going after IoT? This space has yet to see some “real” announcements in terms of the value IoT delivers. And the value that IoT delivers has yet to be truly understood.

Finally, given that POWER8 is supposed to exceed x86’s processing power, will the servers be priced attractively to encourage adoption? IBM states that its new servers will help with “scale-out consolidation”. But remember that Big Data workloads are being processed on scale-out architectures using commodity servers. Is the market at the point where scale-out consolidation is even an issue?

IBM, congratulations on announcing some interesting additions to your Big Data portfolio. But I will wait and see to determine how quickly customers see the value of these new offerings…

Source: IBM Tackles Big Data Challenges with Open Server Innovation Model