The Trouble With OpenStack

Some five years after the launch of OpenStack as a cloud management framework, the platform’s adoption continues to be artificially constrained. While there have been some notable implementations by IT organizations with a lot of internal IT engineering talent at their disposal, mass adoption of OpenStack has yet to occur for a variety of technical and business issues.

At the top of that list is the simple fact that building clouds remains complicated, OpenStack itself still doesn’t scale, and the way the platform is positioned against offerings from VMware may be seriously flawed.

Speaking this week at an OpenStack Silicon Valley event, Randy Bias, vice president of technology for the Emerging Technologies Division of EMC, said configuration issues, in particular, are hampering OpenStack deployments in the field. Given all the compute, storage and networking technologies involved, Bias said there are well over 500 configuration options that can easily conflict with one another. For OpenStack to truly succeed, there’s a crucial need for the OpenStack community to narrow the number of default options available when configuring that platform, Bias said.

Overall, Bias said the adoption of OpenStack as a private cloud computing platform has been stymied because, despite all the marketing hype, there is nothing easy about building, deploying and maintaining a private cloud.

In addition, Bias noted that IT organizations that provide clouds on commodity hardware frequently run into unanticipated engineering issues while the OpenStack platform itself doesn’t yet really scale to meet enterprise IT requirements. Also, there is no real way to test interoperability between different implementations of OpenStack to ensure compatibility across various vendor distributions.

Finally, in terms of positioning OpenStack as an alternative to VMware, Bias observed that OpenStack was never designed to run, for example, SAP applications in an environment that supports synchronous replication, high availability and advanced disaster recovery features. As such, Bias contends that vendors need to do a better job at positioning OpenStack as a platform that is primarily optimized for a new generation of cloud native applications as opposed to engaging in “crazy talk” about replacing VMware, which is owned by a sister company of EMC.

Obviously, many vendors are working on all these OpenStack issues, so, in time, many of these limitations finally may be addressed. For example, Intel earlier this week pumped $100 million into Mirantis, a provider of a distribution of OpenStack, to help fund the engineering work needed to make OpenStack both scale up and out. For its part, Mirantis also has developed a series of appliances that are designed to make OpenStack a little easier to deploy.

Nevertheless, Bias said there is real cause to be concerned about the future viability of OpenStack. The simple fact is that in a world where there are not only alternative management frameworks emerging rapidly, the vendors that govern OpenStack are simply moving too slowly to enable it to become a truly commercially viable entity that solution providers across the channel can deploy in enough volume to build a viable business practice around.

Michael Vizard has been covering IT issues in the enterprise for more than 25 years as an editor and columnist for publications such as InfoWorld, eWEEK, Baseline, CRN, ComputerWorld and Digital Review.