Opinion on IBM’s Direction of OpenSource Application Adoption

July 26th, 2016

Ron Gordon
Director – Power Systems

IBM has made several “product change of direction” announcements over the past several months. They all seem to point to the adoption of OpenSource code as the future direction for several infrastructure and solution applications. The cloud, an automated provisioning area which used to be supported by System Director, has move to an OpenStack implementation now named PowerVC. IBM Cloud Orchestrator is now primarily an IBM GUI, working with a user-supplied OpenStack version.  Smart Cloud Entry became IBM Cloud Manager with OpenStack, and is now IBM Cloud PowerVC Manager, again based on OpenStack.   Just recently, IBM has endorsed the Open Hadoop and Spark versions, and it appears they will slowly phase out the BigInsights product and family. IBM has also been strongly endorsing Data Base portfolio offerings.  Another adoption is PowerKVM, which is a POWER technology-enhanced version of KVM, for virtualization of Linux environments.

My opinion is that this a very wise move.  It is good for both customers and IBM, as well as for Mainline as a consultant and advisor to our customers. OpenSource and the open community are driving advancements at a pace that has previously not been possible. Historically, ISVs, application developers, and software companies developed with a relatively small team, and functional advancements were based on the defined direction the organization wanted to go, and usually had annual updates only.  This has all changed, with OpenSource and community-driven development. Many more developers contribute rapidly to the functionality, based on the market and community needs. The validity and strength of the code is driven by the community, with great pride in achievement and the desire for needed functionality. Besides the cost advantages of OpenSource, mostly controlled and licensed under the GPUL, users can see a long life expectancy and growth in OpenSource.

As many customers see support as a critical element of software, and cannot risk software with only community support, which may not be timely nor complete, we see many applications being supported with what is usually called an “enterprise version.”  A company (i.e., Cloudera, Hortonworks or Redis Labs) uses OpenSource code as a base, embraces its growth, and usually provides some optimization for certain architectures, support, and compiled versions, for easier implementations. The enterprise versions usually come with a price and a value-add, over the GPUL licensed OpenSource code.

What about operating system support? Well, most all of the OpenSource code coming out is using Linux as the underlying Operating System. Why? Because Linux is ubiquitous, and so is OpenSource. This ensures application developers that their code is available on almost any system architecture. This is symbiotic by design, in my opinion. Customers can choose OpenSource Apache and run it on OpenSource Linux with open extensions, and provide a very low cost, highly effective solution.

Mainline