Building the convincing PaaS business case

Kurt Marko Profile picture for user kmarko January 16, 2018
Facilitating the rapid response to newly exposed security holes is a compelling argument for taking a systematic, automated approach to applications and IT systems. However it’s but one advantage of many.


Cloud revs

As I previously detailed, the pervasive Meltdown and Spectre hardware security flaws:

...should be a wakeup call for organizations to build automation into their infrastructure to minimize or eliminate the overhead of software updates,

proceeding to extol the virtues of PaaS as a process automator.

However, inserting an abstraction layer between infrastructure and its users, namely developers and IT operations (or DevOps teams in progressive organizations) yields many other benefits. Rapid deployment of security patches is one of the more mundane, if not significant, benefits of a PaaS stack and associated processes. While easy to enumerate, as we’ll see, the business rewards of PaaS are harder to quantify, although according to a few studies, they are significant.

Qualitative benefits

As both anecdotal evidence and formal studies from consultants like Forrester, IDC and KPMG demonstrate, incorporating PaaS into software development and release processes yields several significant business benefits, including:

  • Faster release cycles, sometimes requiring as little as one-tenth the time to develop and provision a new application or IT service.
  • Concomitantly increased revenue via faster product development and delivery. Furthermore, by standardizing development and implementation processes, PaaS steepens the curve of cost-reduction over time, thus improving margins earlier in the product cycle.
  • Better software quality that requires fewer patches and less downtime by enforcing structured processes and code reuse easier module integration, testing and revision management.
  • Improved customer satisfaction and service via consistency across application and infrastructure implementations that leads to fewer bugs, and faster problem resolution. Consequently, support staff deals with fewer calls, resolves them more quickly, at lower cost, more often remotely without the need for on-site, in-person visit.

The combination of increased developer productivity and lower administrative overhead changes the way software-centric organizations do business by sharpening the focus on products, not process, as evidenced by this quote from the lead developer at the Kahn Academy speaking about its experience with Google App Engine,

We don’t need a system administrator or anyone dedicated to deploying our app, so 99 percent of our time is spent working on our application. — It’s nice to have an easy deployment process since we deploy the site on average at least once a day, [and] sometimes up to nine or 10 times [a day].

Financial rewards

The problem with cost models and ROI analyses are that they are hard to generalize and inherently easy to game by making assumptions and cherry picking data that tilt towards a preferred conclusion.

The most meaningful models are those customized for a particular organization and business situation. Given the ability to bend the models to a desired outcome, financial benefits are even more suspect when the analysis is done by vendor-sponsored consultants. Sadly, when it comes to relatively new and evolving technologies like PaaS, these are the only ones publicly available since any organization that goes to the trouble of building a detailed financial analysis using data from private metrics isn’t likely to openly share it with competitors. Nevertheless, we can assess such studies with a skeptical eye using logic and common sense to at least validate the qualitative claims.

The most detailed financial analyses of PaaS that I've seen published are both vendor-sponsored, one by Azure, the other by Pivotal, but each conducted by Forrester, an organization with people that know their way around spreadsheets.

They each attempt to quantify benefits such as lower IT administration overhead, avoided and reduced IT resource (equipment, software license) costs, faster deployments and time-to-market, reduced costs for software testing and support, and the business opportunities enabled by freeing up IT and developer resources. The Pivotal paper calculates the following financial benefits

  • Operations productivity through an automated toolchain and deployments: +20% in year 1, 40% in year 2 (25% incremental) and 60% in year 3 (33% incremental),
  • Developer productivity again through automation, faster access to development and testing resources and faster code integration: +25% in year 1, 35% in year 2 (13% incremental), 50% (23% incremental yr 3)
  • Platform costs via system consolidation through higher resource utilization and fewer required non-production environments for dev/test: 30% in year 1, 50% in year 2 (29% incremental), 80% in year 3(a credulity-straining 60% incremental)

The Azure-sponsored paper that produced five-year cost estimates for a representative organization yielded an incredible and unrealistically high 466% ROI resulting from,

An 80% reduction in IT administration time required to manage apps deployed on the platform, a 25-hour average reduction in development  and testing time required to develop or update Azure PaaS applications, and a 50% reduction in time required to help deploy a new application solution to a client.

The analysis concludes that PaaS surpasses IaaS by eliminating time spent on infrastructure setup, management and troubleshooting, allowing adopters more time for product development. Similar to the Kahn Academy experience, the paper quotes the founder of an IT services firm extolling the benefits,

Developing our app without PaaS? Well, we wouldn’t have done it. Frankly, the extra time and resources required would have wiped out any profit that we have generated.

Ultimately, the business case for any product or technology rests on its market success or failure, and here, the case for PaaS is gaining traction.

The smallest cloud segments, one of the fastest growing

Market estimates are subject to the same sort of statistical torture as ROI models, so caveat emptor. However several estimates demonstrate that the PaaS market, while small today, is expanding. Gartner recently pegged PaaS revenue at about one-third that for IaaS, but projected to grow more than 23% annually through 2020. A CIO survey and analysis from KPMG was more optimistic, noting that (emphasis added),

Cloud is poised for mainstream adoption and even more rapid growth. While investments in SaaS and IaaS will continue to grow, investment in PaaS will almost double its growth rate as CIOs adopt cloud computing as their strategic development and production platform for both new applications and services, and as they begin the process of paying down technical debt and migrating or modernizing their legacy IT estate.

The KPMG report also cites estimates from Wikibon showing PaaS emerging as a multi-billion dollar business over the next decade. Given how the lines between IaaS and PaaS continue to blur just as organizations are starting to see the value of more abstract, structured application services, I believe that such estimates are too low. Indeed, KPMG notes the emergence of industry- or application-specific PaaS like GE Predix, as another catalyst for growth, writing,

Adding to complexity is the recent emergence of industry cloud platforms. These are typically positioned as a specialized PaaS offering with API interfaces to a catalog of cloud-based services that developers can use to accelerate the creation of apps and deliver more value to their customers.

My take

IaaS and SaaS have benefited from being conceptually similar to familiar IT product categories, namely virtual infrastructure and browser-based applications respectively.

Each could displace existing services with minimal disruption. In contrast, PaaS entails rethinking core software development and IT management processes, often under the rubric of DevOps. The PaaS case requires a level of strategic and cultural buy-in that normally hinders rapid adoption.

Nevertheless, if PaaS truly delivers a better mousetrap, more and more organizations will take the plunge, which is precisely what’s been happening. Look no further than the broad participation in and enthusiasm about Cloud Foundry as I detailed after attending its last US Summit.

PaaS is a manifestation of the maturation and ‘mechanization’ of IT services and software development in which infrastructure implementations are rentable commodities and competitive differentiation lies in higher-value digital services. By streamlining the production and maintenance of such services without compromising creative innovation, PaaS is a natural evolution of software methodologies that exploit the advantages of cloud services. Look for PaaS to be an increasingly competitive battleground among major cloud players and software platform vendors this year and beyond.

A grey colored placeholder image