There was a time when there were several attractive private cloud and cloud stack options, but as the market has...
matured, the name on everyone's lips is OpenStack. As recently as a year or two ago, there were at least five evenly matched OpenStack alternatives: CloudStack, Eucalyptus, Nimbus, OpenNebula and vCloud. Today, cloud pros wonder if there is any real competition left.
What's made the difference for OpenStack has been vendor support. Open source software is popular, but most enterprises want and need support for their cloud platforms. Computer vendors HP, Dell and IBM, and network vendor Cisco see OpenStack as a way of entering the cloud business, while profiting from professional service opportunities and symbiosis between hardware products and cloud. OpenStack gives those vendors a public relations jolt, as well as credibility. That type of impact means OpenStack has leap-frogged its competitors.
The victory celebration, however, may be premature. There are three market forces that could unseat OpenStack and tear the whole cloud stack software market open.
Threats from the reigning public cloud king, Amazon
The first threat to OpenStack's reign of supremacy is competition from the cloud giant, Amazon Web Services (AWS). Currently, AWS does not offer private cloud software, but it could feel pressure if OpenStack evolves into hybrids that favor public cloud services. AWS could then promote Amazon-friendly private cloud stacks or make its own cloud software available for private and public cloud use.
However, Amazon may not choose to be an OpenStack competitor for open source cloud software. The provider should fear the possibility that a startup would select an AWS-supported cloud stack to compete with its own public cloud. The question is whether a service provider, cable or computer company would take that competitive step. If so, Amazon would not be able to control how open source software was used, even if it promoted and contributed to it.
AWS could enter the market themselves and make a version of their own software available for the private cloud. That would make Amazon a leader in cloud software, if they avoid open source licensing.
A focus on cloud-independent deployment tools
The second threat to OpenStack is a focus on deployment tools that are cloud-independent. Most complex cloud applications are deployed using DevOps or orchestration tools, while most commercial and open source tools already support multiple cloud software stacks. As tools become more refined, users will view cloud only as the bottom of an application lifecycle management process. The tools become what users commit their cloud deployments to. All x86 servers appear similar under the same operating system and middleware platform. The same is true for clouds under deployment and orchestration toolkits.
The most significant development in deploying applications and components in the cloud is the emerging Topology and Orchestration Specification for Cloud Applications (TOSCA) from OASIS. A TOSCA-based approach to DevOps creates a uniform model to describe application deployment and management -- and any cloud stack could then be supported with the proper TOSCA plans. Movement among clouds becomes so easy that no cloud stack could lock in users. However, TOSCA may take time to gain acceptance and may never fully define cloud orchestration practices. In that case, it would only fragment the space further and reduce the threat to OpenStack from this quarter.
Cloud Foundry: The poster child against OpenStack
The third threat to OpenStack comes from IaaS intransigence. OpenStack specifically targets offering infrastructure as a service (Iaas). There are strong signals that IaaS is not the optimum cloud service strategy for public cloud providers, and that software as a service is better for users and providers. Efforts to expand OpenStack into platform as a service (PaaS), however, have been met with resistance from within the OpenStack community.
Cloud Foundry is the poster child for an OpenStack threat. Most companies interested in cloud computing know that uniform sets of OS and middleware tools to support application mix simplify operationalizing private cloud. Public cloud providers know that add-ons to IaaS raise profits by displacing buyer IT costs, creating a win-win. Cloud software that elevates the cloud service to PaaS makes the underlying IaaS choice less important and harder to differentiate IaaS cloud stacks. IBM and HP demonstrate the power and potential OpenStack risk that PaaS could present.
Any of these risks have a modest chance of becoming critical in the near term, but it's likely that at least one of them will emerge as a crucial factor in the cloud software market down the road. In particular, it appears that the movement up the cloud stack toward PaaS and the increased need for DevOps and orchestration tools will slowly submerge many of the current cloud software issues. OpenStack might be a winner in an increasingly invisible market race.
About the author:
Tom Nolle is president of CIMI Corp., a strategic consulting firm specializing in telecommunications and data communications since 1982.
Cloud's future peppered with PaaS, open source
HP gets comfortable with OpenStack for open source push