BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
There were a lot of important cloud developments in 2016, and many of the most important ones came in the platform-as-a-service space. This year, we saw the emergence of PaaS as a boundary space between basic infrastructure as a service and advanced cloud services of all types. PaaS is more than a middleware-equipped cloud, more than Azure. Now, the PaaS business model is a totally new software development platform and some of the cloud services are complete applications.
Moving PaaS beyond IaaS
We tend to think of PaaS as an alternative to IaaS, where systems software is included along with hosting in an overall cloud package. A better definition of PaaS would be a cloud service where developers incorporate cloud-hosted software logic into their programs; that logic is called platform. That platform notion evolved significantly in 2016 -- both in terms of the features it offers and in terms of how it's driving cloud adoption.
IaaS is a universal service where you can host almost anything in a virtual machine (VM). The problem is if all the cloud can do is offer hosted virtual machines, then its benefits are limited and so is its adoption. The least PaaS can do is displace more cost by offering more features, and that builds the value of the cloud. That's the role of PaaS today. PaaS features can actually reshape how applications are developed, to the point where some applications that now run in the cloud wouldn't really run anywhere else.
The original PaaS business model of comprehensive middleware tools combined with container or VM hosting is still a perfectly fine PaaS business model. It's offered both for the Microsoft community (.NET and other Windows Server middleware) and the Java community. Both Microsoft middleware and Java have become more inclusive of the other environment (you can call .NET components from Java and use Java components in .NET applications). We now find it easier to adopt PaaS, irrespective of current data center architecture. Since Java is the language of the web, this cross-connection with Azure and .NET opens web concepts like microservices to Azure and even Windows Server users. The feature framework of PaaS is then expanded.
Traditional PaaS and IaaS added another dimension in 2016. Both Amazon and Microsoft enlarged their repertoire of cloud APIs. Branching out from basic database services or development/deployment tools to specialized services designed to support new programming models (functional programming) and specific new technologies like mobile broadband and the internet of things (IoT). These new APIs offer cloud providers new sources of revenue and also let users build applications using cloud-specific tools, which enhances the benefit of the cloud. Through the addition of these cloud-hosted APIs, IaaS and the old PaaS model are converging on a kind of functions-as-a-service approach.
Functions and APIs are development tools, so all of that has shifted the cloud toward a new model of a compute platform, complete with virtual operating system, middleware and other tools. Because of that shift, enterprises are now developing applications that would run with difficulty, if at all, in the old data center model of computing. In 2016, this started moving the cloud as a whole toward a model of distributed computing and not just a model of hosting.
PaaS and future opportunities
Mobility and IoT are two areas that enterprises prioritize in terms of future technology planning. There's no coincidence that both of these areas are focus points for cloud-service APIs that convert IaaS into something PaaS-like and move PaaS closer to offering whole applications. Most companies have only begun to invest in mobile empowerment and most haven't started exploiting IoT more broadly. The availability of specialized cloud services, a special PaaS business model, means that these critical technologies might not have to be moved to the cloud because they'd start there and never leave.
Amazon's recent announcement of general availability of its QuickSight populist analytics and data visualization tool is an example of a PaaS feature set being married to a user-friendly front end and packaged as a service. Not only does this demonstrate the evolution of IaaS toward PaaS through the addition of web services, it also shows how PaaS could evolve toward SaaS.
A more direct example of that can be found in Oracle's cloud strategy. Oracle characterizes its Application Cloud as software as a service (SaaS). It is because the APIs it offers link not to logical components or tools, but to business applications. However, the primary benefit of the Application Cloud is that it's SaaS for a developer and a platform to facilitate SaaS integration with existing or even evolving data center or traditional IaaS/PaaS applications.
A sum of the changes in PaaS have already impacted application development and IT operations. DevOps, once a kind of geeky enhancement to operating system scripting, is becoming the central tool in managing application lifecycles where components are hosted in the cloud or the data center and are developed by users or deployed by cloud providers as part of PaaS. All the popular DevOps tools are getting more sophisticated. New competitors are announcing products and the goal of full automation of application deployment and redeployment is reachable, perhaps for the first time.
PaaS model and your business
From a business perspective, the PaaS changes are easing profit pressure on cloud providers -- that is, by providing new opportunities for feature differentiation and by opening new applications to the cloud, expanding the total market. Cloud spending, which has yet to reach even 10% of IT spending, has been benefit-constrained. The evolution of PaaS has broken that barrier now, and we can see how the cloud model could completely absorb business IT, pulling public providers and private data centers into a common architecture.
The year 2016 proved, in a value sense, that all of the cloud and data center applications are really elements in a global, systemic PaaS, whose specific elements differ only in the number and nature of platform services. This could be the first year when we could say that, through an enhanced PaaS business model, the cloud is truly everywhere.
Learn about PaaS cloud app security
Learn how to improve portability and flexibility in a PaaS model
Things to consider while choosing IaaS vendors