PERSPECTIVES

API Modernization: A Novel Approach Pays Dividends

2022-03-10


Migrating from integration-platform-as-a-service environments to AWS public cloud can enable greater flexibility, faster time-to-market, better user experience and lower cost. Interestingly, the approach we used to assist a global data provider can be applied to nearly any company seeking to move from IPaaS to a public cloud.

Application integration is one of the most important capabilities for the would-be modern enterprise. Without it, businesses can never bridge old with new technology and process, and fully utilize disparate systems, services and data. But as with most technology adventures, it is easier said than done.

The introduction of integration platform as a service (IPaaS) environments has, inarguably, been a major advancement enabling organizations to replace manually intensive application integration activities by essentially renting integration services from third parties, namely cloud providers.

Greater affordability and scalability

Over time, however, IPaaS has become expensive, especially with the advent of infrastructure-as-a-service (IaaS) offerings on the public cloud. Providers of public cloud services — the so-called “hyper-scalers” such as Amazon Web Services (AWS), Microsoft and Google — have the theoretical ability to scale infinitely. For organizations seeking to modernize their IT infrastructure, working with hyper-scalers presents a more attractive option by offering greater affordability, greater scalability (both up and down) and agility for the business to respond to changing conditions.

Not surprisingly, many IPaaS customers are looking to move to aforementioned cloud platforms. Making the move is a complicated technical undertaking with few available solutions. The missing piece: A reusable application that includes modern application programming interface (API) capability to help accelerate integration services between logical network endpoints.

Last year, we set out to build a reusable framework that would enable a global provider of business information to more effectively and quickly transition from IPaaS to IaaS.

Move from IPaaS to IaaS

Our customer, based in Europe, is among the world’s largest providers of business credit reports, serving 365 million businesses worldwide. As an information provider, its business depends wholly on integrating data from a vast multitude of sources, so timely and cost-effective application integration is paramount. However, it had a major challenge: Facing an upcoming IPaaS renewal date, it wanted to jettison IPaaS and latch onto IaaS within a six-month window.

We proposed and then utilized a unique approach that required the client to migrate use cases rather than services. We then helped the company move onto AWS, a platform that will deliver much greater operational agility and cost efficiency. This approach, we have found, can help companies across industries move from virtually any IPaaS platform onto the public cloud.

A focus on experiences, not services

Thankfully, we completed the transition for this client within the six-month timeframe. Our client is now better positioned to make changes to its business processes that will drive speed to market. It also expects to achieve major benefits in terms of greater performance that will drive better user experience and more efficiency, which will improve IT staff and user experience.

The company does not yet have a clear picture of the exact cost benefits of its move to AWS, but it expects the cost reductions in the range of 25% annually over the IPaaS solution. And the company is already seeing a 25% to 30% reduction in the time needed to make system changes.

Historically, a company like this would migrate service by service at the experience layer, process layer and system layer. But that can be problematic. In our experience, almost one-third of endpoints are typically deprecated (no one uses them), but they are still present. Taking a service-by-service migration approach would mean moving endpoints that are defunct, wasting a lot of time (and therefore money).

Focus on use cases for migration

A better way is to migrate use case by use case. Use cases might range from such operational functions as “onboard new customer” to those related to the core business, such as “list all retail companies over $1 billion in revenues” or “list the top executives of XYZ company.” Focusing on use cases is much more efficient than migrating services because if there is a function that is not used in any of the use cases, then it will not be migrated to the new system.

This method, we have found, helps identify which endpoint is being used. An additional advantage: Migrated services cannot be tested in isolation, whereas migrated use cases can. This approach enables faster yet, at the same time, more gradual adoption of the new API delivery platform.

While hyper-scalers provide API delivery tools and tackles, they do not offer the ways and means for migrating integration flows from IPaaS platforms to their cloud environments. Moreover, the phased release of migrated endpoints enables migration challenges to be mitigated more manageably. This is in keeping with a “fail-fast” design philosophy of today’s minimum viable product (MVP) movement, which is critical in a time-sensitive modernization project. (Here’s a deeper dive into the technical aspects of the migration and our partnership with AWS.)

Smooth pathway from IPaaS to IaaS

Our work with this client reveals how a use case migration approach enables a smooth pathway from IPaaS to IaaS. In fact, we learned that this novel approach can be applied easily to companies across industry segments that are seeking to decrease IT operational costs and speed time to value of new business capabilities via an IPaaS to hyper-scaler cloud migration, regardless of the platforms involved.

This article was written by Dipanjan Sengupta, a Senior Director in Cognizant’s Digital Engineering Practice.

For additional insights, visit the Digital Engineering section of our website or contact us.