Pivotal, the Big Data spinoff from EMC and VMWare, has announced today its "Pivotal Big Data Suite," which it refers to as the industry's first Big Data Mega Bundle. The suite, available on a subscription basis, includes most of the data-related components in Pivotal's portfolio. These include:
- Greenplum Database, Pivotal's Massively Parallel Processing (MPP) relational data warehouse
- Pivotal HD, the company's Hadoop distribution
- HAWQ, a SQL-on-Hadoop engine that integrates Greenplum technology into Pivotal HD
- GemFire, a distributed, in-memory transactional database
- GemFire XD, a Hadoop-based in-memory database for Java workloads
- SQLFire, a "NewSQL" (relational, in-memory, scale-out) database
Big Data Suite subscriptions are priced on a per-core basis, for two- or three-year terms. The company hasn't specified actual subscription costs but describes the suite as being "priced aggressively."
Per-core pricing means that customers are not taxed for storing large amounts of data, but only by the amount of processing power they require. Moreover, Pivotal says that the Hadoop component of the subscription, Pivotal HD, is provided under an "umlimited" subscription. Exactly what that means in the context of per-core pricing isn't clear.
All you can eat
Pivotal's strategy here is an interesting one. The company isn't championing one big data architecture over another, or advocating for several architectures while charging for each one separately. Instead, the company's providing a universal subscription, establishing a sort of big data flat tax.
Microsoft alumnus Paul Maritz, Pivotal's CEO, said "With the Pivotal Big Data Suite, we are taking the lead for the industry by removing the technical barriers to data off the plates of our customers and [letting them] move to a world where the choice isn’t about Hadoop or a SQL database, in-memory or real-time processing, but to efficiency and value."
More is better?
A more cynical view might be that Pivotal, with its Greenplum and HAWQ foundation, accompanied by a selection of more obscure and less integrated technologies, has found a way to rationalize its stack. Suddenly the more minor components become value-adds rather than curious appendages that might be hard to license on their own.
On the other hand, there is little question that today's Big Data landscape involves a combination of Hadoop, MPP, and in-memory technologies. If Pivotal wants to win the multi-front war against Hadoop-focused competitors like Cloudera, and Enterprise competitors like Teradata, IBM or HP, then maybe a catch-all subscription isn't such a crazy idea.