A key draw of the cloud is the way it simplifies software licensing. Instead of perpetual licenses, you bought subscriptions. While the spotlight was on the fact that cloud subscriptions could come out of operating, rather than capital budgets, subscriptions carried another hidden advantage. If you didn't need the software or service, you weren't burdened with a sunk cost; you could simply not renew the subscription.
Pressure for simplification has bled over to more traditional on-premise software markets. While we predict that cloud will become the default choice for new big data workloads by next year, few enterprises are going to ring up Amazon to dispatch Snowmobiles to completely hollow out their data centers.
For instance, Teradata recently revamped its bundling with a common Teradata Everywhere pricing scheme based on a TCore metric that, believe it or not, is actually much simpler compared to what it replaced. In turn, Teradata now bundles tools with its platform that used to be priced a la carte. Oracle has simplified pricing for many of its PaaS and SaaS offerings in its cloud, with the brunt of contracts being completely automated. Cloudera, which used to package its platform based on technology features, now bases pricing on use case.
IBM is the latest to catch the bug for refactoring the bundling of its analytic platforms into three broad bundles covering database; data governance and integration; and data science. Within each of these bundles, you can mix and match
platforms components based on buying units of "FlexPoints," that are based on virtual processor cores. The operable notion is that, in contrast to traditional, perpetual licensing of commercial databases and analytics tools, you will not be stuck with a stranded investment should your business needs change in midstream.
So for instance, the database platforms give you a choice of the Db2 operational database; Db2 Warehouse; Db2 Event Store (an in-memory Parquet file store designed for fast ingest of streaming data); and Db2 BigSQL (which provides federated query to Hadoop). During the life of your agreement, you can swap one out and replace it with the other. However, the more likely scenario is that you've bought a certain amount of Db2 Data Warehouse, but then find new use cases for real-time processing; you can then reduce the number of FlexPoints for the data warehouse to cover growing the footprint of Db2 Event Store.
As for the other packages, the governance and integration bundle covers Information Server for ETL; BigIntegrate (ETL on Hadoop), BigQuality, BigMatch, Information Governance Catalog, Master Data Management, Info Lifecycle Governance, and several related offerings. The data science bundle offers a choice of Data Science Experience, SPSS Modeler, Cognos Analytics, Watson Explorer, and several related products.
For now, the focus of the new bundles is to provide more flexibility in mixing and matching data platforms, integration, governance, and advanced analytics offerings. It is not about refactoring or reachitecting products per se, but that doesn't mean that IBM has sat still on that front. For instance, the addition of a common machine learning engine between IBM Master Data Management and Information Governance Catalog automates data matching and metadata generation.
TechRepublic: IBM Watson: The smart person's guide
The new FlexPoints mix and match program does not at this point extend to IBM's open sourced big data products, such as the Hortonworks Data Platform that IBM resells under an OEM agreement. The same goes for other open source data platforms that IBM currently hosts on its cloud such as PostgreSQL. Admittedly, the subscription pricing model for open source products and cloud services is obviously different, and in fact the subscription feature itself provides a similar degree of flexibility at the end of any annual contract. That said, there's no reason why IBM couldn't offer similar mix and match flexibility in an open source bundle, especially with managed services in its cloud.