Skip to main content

Navigating the PaaS Maze

By June 18, 2012Article

In this article, I will explain platform as a service (PaaS) as we know it today, its connection to the past and the reason why current PaaS offerings meet the necessary condition but are not sufficient. I will also explain the difference between iPaaS and aPaaS and cover some prominent vendors in those respective categories. In addition, I will delve deeper into meta-data aPaaS and general purpose programming aPaaS and provide key points that help enterprises choose the right aPaaS.
What “platform” meant pre-cloud
In the pre-cloud, era there were no platforms; there were only products. The IT departments in big enterprises had “enterprise architecture groups” that defined the platform for their enterprise. They hand-crafted a platform definition standard for their enterprise by combining various products like app servers, web servers, databases, middleware, integration servers, portal servers, workflow engines and so on.
This was still a definition in a blueprint form, “not an installable platform.” That is the only way they could manage the IT software infrastructure of very large multi-country, multi-location enterprises. Every time a new project sprang up, the project team created an instance of the platform as per their blueprint. Every large enterprise had its own blueprint platforms. They did a lot of heavy lifting not directly connected to their business outcomes.

Platform on the cloud
Different technology pieces required for a platform have now come together as a pre-integrated, ready-made stack. In one single click, version upgrades of each component in a blueprint platform and related change-management nightmare of the pre-cloud era all have disappeared.
All these technological advancements coupled with delivery model and business model innovation by providing it “as a service” and “pay by the drink” brings in a disruption. This presents an opportunity to move away from large capital technology expenses and align technology costs directly with business growth.
Innovative enterprise IT leaders and the “fast followers” have adopted this trend (see this video address of Casey CIO of GSA). Even government agencies, which are generally slow in adopting new technologies, are running fast.

Platform of the future — beyond technology
My personal view is this pre-integrated stack of technology that we call “platform” should have happened around the 2003-05 timeframe. But at that time it lacked tailwind, something we have now in the form of cloud. Hence the PaaS solutions we know of today in my view are only a necessary condition and not sufficient to really help large enterprises with where they need to go.
The “platform-of-the-future” solution goes beyond technology. The enterprise architecture teams in large companies need to move on by delegating platform definition, blueprinting and construction to platform vendors and move into “platform governance.” They need to acknowledge that the people closer to the problem not only have greater understanding of the business problem but they also have the quickest and easiest solutions. They also know reasonable compromises they can make in their requirements and get things working today instead of a few months later.

 

The true “platform of the future” goes beyond the technology stack. It is a platform that drives business innovation to the core and leverages the knowledge that lies in the fringes of enterprises. It is a platform of agility. It is the platform of rapid prototyping. It is the “collaborative platform” where lines of business and enterprise IT come together in bringing to life innovation solutions for their enterprise.
Platform-as-a-Service types
Before discussing the players in the PaaS landscape, it is important to understand the two broad categories of platforms: “iPaaS” and “aPaaS.” iPaaS is an integration platform as a service; aPaaS is application platform as a service. These are not new concepts. They roughly map to the integration middleware and application middleware in the pre-cloud paradigm

iPaaS has integration orchestration, a set of pre-built connectors to standard cloud services and popular on-premise enterprise software like SAP and Oracle Apps.
The good news for enterprises with respect to the iPaaS category is that it is easier to make vendor choices in this category compared to aPaaS. There are only three clear choices: two commercial vendors (Boomi and CastIron) and one open source option (Mule iOn).
Application platform as a service
Unlike iPaaS, the aPaaS landscape is much more crowded. Also, enterprises have to watch out as there is lot of “cloud washing” in this category. In cloud washing, existing vendors offering on-premise solutions quickly rebrand them as “cloud” offerings without making the requisite changes to the product.

Within aPaaS, there are two distinct categories of users/buyers:

  1. Independent Software Vendors” (ISVs) that build SaaS using a PaaS solution
  2. Enterprises that need to build applications that are specific to their needs and differentiated from their competitors.

The business goals and motivations of these two buyer groups are in many ways completely different as you can see from the diagram below.

The motivations of “enterprise IT” are:

  • Productivity – Enterprise IT leaders want their teams to have a highly productive development platform instead of figuring out the nuts and bolts.
  • Time to market – As we all know, a line of business always has “This is needed yesterday” type demands. Hence, time to market to stay ahead of their competition is a primary motivator.
  • Governance and abstraction – Every year there are additional governance needs. Enterprise IT knows the easiest way to address governance is by reducing the moving parts. Abstraction is the mantra here.
  • Means to an end – Enterprises are into their main business of making cars or digging of oil or providing healthcare. They don’t make money selling software. For them, software is a means to an end. Hence the platform of the future needs to address this “means to an end” goal.

In contrast, the motivations of ISVs are:

  • Control – ISVs look for fine-grain control to be able to choose all the moving parts that go into the platform. This is one of the reasons ISVs generally tend to take the Infrastructure-as-a-Service (IaaS) route instead of the PaaS route.
  • Multi-tenancy – Ability to service multiple customers on the same infrastructure and take advantage of economy of scale and load offsetting.
  • Coding and granularity – This is an additional dimension of control, but in a different plane. For example, an ISV may have legitimate reasons for not using ORM (Object Relational Mapping) for their SaaS application and prefer instead to do something on their own. Their SaaS application might demand that.
  • Breadwinner – ISVs’ SaaS offerings are their breadwinner, revenue generator and their IP. They are protective and possessive about these offerings, and this drives lot of their decision making in the kind of platforms they choose.

Based on the above motivations, aPaaS players can be classified into “meta-data aPaaS” and “programming aPaaS.” Meta-data aPaaS doesn’t mean that there is no possibility of coding. It is really the 80-20 rule; 80 percent is meta-data modeling, abstraction driven and APIs.
With this classification of meta-data aPaaS and programming aPaaS, a better picture of the vendor landscape in aPaaS emerges.

Application PaaS selection criteria
When choosing an aPaaS, enterprises need to look for an aPaaS that:

  • Creates and manages business objects and not code
  • Allows resorting to code for extensions
  • Defines business processes through a workflow engine
  • Implements logic using business rules instead of burying them in code
  • Includes drag-and-drop UI creation for easy change management
  • Ensures all business objects are SOA compliant through REST and/or SOAP
  • Provides multi-device support; apps should be accessible from desktops and mobile devices including laptops, tablets and smartphones
  • Provides a library of template applications to get off the ground quickly
  • Enables the ability to run on-premise and in private and public clouds. This is a very important criterion when choosing a “platform of the future,” especially in an enterprise context.

To summarize, enterprises have to pick one integration platform as a service (iPaaS). As I articulated above, iPaaS selection is likely to be less complicated than aPaas. Choosing the right aPaaS can be the difference between winning and losing. History has proved time and again that large established incumbents rarely offer paradigm shifts. It is the Davids who beat the Goliaths. Enterprises need to look beyond their comfort zones into new players that offer aPaaS that could put them ahead of the competition.
Suresh Sambandam is the founder and CEO of OrangeScape Platform as a Service, one of the global 10 PaaS companies featured in research reports of Gartner and Forrester with marquee enterprise customers like Unilever, Pfizer, AstraZeneca, United Biscuits, India Infoline, Fullerton and more. OrangeScape is the only cross cloud platform that supports private cloud deployments through Eucalyptus and popular public clouds like Google App Engine, Windows Azure, IBM Smart Cloud and Amazon AWS.

Copy link
Powered by Social Snap