To put the discussion of emerging cloud standards into perspective, it makes sense to step back for a moment and review basic terminology. For the most part there is consensus that cloud architectures roughly divide into three logical abstraction layers: Software-as-a-Service (SaaS), Platform-as-a-Service (PaaS), and Information-as-a-Service (IaaS). This is often referred to as the Cloud Service Stack Model.
While I would not go so far as to say that all the boundaries between each of the three layers have been completely standardized yet, there is, for the most part, agreement in the cloud community. Standards beyond the three simple divisions are fuzzy at best. There is much room for the creation of viable standards within, between, and among each of the cloud layers.
Starting from the most abstract or top layer, SaaS is the public face of the cloud, encompassing applications such as Salesforce, Workday, Mozy and other tools used by millions of businesses and consumers. When applications are built in a private cloud within the enterprise, these are often referred to as cloud-ready applications. Standards at this layer of the stack are scarce. The SaaS vendors have little or no motivation for tying their products into a standard ecosystem.
On the other hand, there is growing pressure from users to create interoperability standards. In an ideal world, not only would this allow consumers to migrate between SaaS product offerings as their requirements change, it would also allow businesses to create mixed portfolios of SaaS and internal cloud-ready applications. Cloud data transfer standards also would reduce customer worry about the very real threat of vendor lock-in. Hopefully no SaaS vendor is so arrogant as to think that his solution is the only one his customers will be using.
Moving down the stack, PaaS includes the development environment and the operations management tools that support applications. Azure offered by Microsoft and Cloud Foundry acquired by VMware are examples of pure PaaS platforms. At the PaaS layer, cloud standards allow companies to create applications that work on a variety of different platforms without the need for costly ports or application refactoring. This layer is so immature that the standards are non-existent, although some people are arguing that Cloud Foundry has the best chance of becoming the de facto standard. I would expect there to be a big push from both vendors and consumers for PaaS standards in the coming year.
Finally, the most conceptually concrete layer is IaaS, which includes the data center, physical systems, networks (sometimes referred to as ping, pipe and power), hypervisors, orchestration tools, and operations systems. IaaS layer standards open up the possibilities for wider adoption of cloud architectures in the enterprise setting by reducing complexity and enabling companies to take advantage of operational best practices and widely understood use cases.
Anyone who has attempted to build even a simple cloud in the enterprise knows just how difficult it can be. The growing popularity of the cloud-in-a-box hardware offerings such as Vblock from VCE and NetApp’s FlexPod attests to the need for good IaaS cloud standards. I suspect that these are a very long way from crystalizing just because most clouds are standalone systems.
Until there is more enterprise pressure to build more interoperability between clouds, most companies will make do with cloud broker services such as enStatus and Rightscale to fill the gap.
Beth Cohen is a cloud strategist for Verizon, helping to develop cutting-edge products for the next generation. Previously, Beth was president of Luth Computer Specialists, an independent consultancy specializing in cloud-focused solutions to help enterprises leverage the efficiencies of cloud architectures and technologies, a senior cloud architect with Cloud Technology Partners, and the director of engineering IT for BBN Corporation, where she was involved with the initial development of the Internet and worked on some of the hottest networking and web technology protocols in their infancy.