Infosys TechCompass

Frameworks and stacks

Trend 5: Focus shifts from languages and frameworks to stacks

With the evolving microservices landscape, more frameworks, messaging systems, transport layers, data serialization formats, APIs, etc., will emerge. Since these components constitute a whole stack of offerings, developers are less likely to choose individual languages and frameworks for their application development.

Earlier, the microservices component landscape offered few choices. Developers were forced to select individual languages and frameworks; web application stacks such as LAMP, WISA, MEAN, and PERN; and Netflix OSS microservices stacks. Some of these stacks are now obsolete, while others try to catch up with required augmentation with newer components to work properly. However, with today's ultralow latency, highly efficient data serialization, and API querying options, these stacks will become more powerful.

A large financial institution, in partnership with Infosys, developed the GRAND stack framework to modernize its payment services technology. This stack uses native end-to-end synergies in place of individual programming languages.

Frameworks and stacks

Trend 6: Polyglot frameworks gain prominence

Modern Java frameworks that offer fast throughput and nominal startup time (e.g., Quarkus, Micronaut, and Helidon) have become instrumental in robust microservice and serverless application builds. These frameworks support Amazon Web Services (AWS) Lambda, and Azure Functions, as well as non-blocking reactive styles of programming and declarative types.

Earlier, the absence of appropriate dependency injection standards, JVM-related limitation on the modules and cloud-native features drove the community/enterprise to use Spring as a de facto standard of development despite Spring lacking the memory efficiency.

New frameworks are polyglot in nature, with serverless extensions and Kubernetes support. While polyglot VM offers ahead-of-time compilation, the frameworks support compile-time dependency injection and greatly enhance the developer experience and runtime performance.

Infosys has used Reactive Frameworks in several engagements. It is now exploring the use of Quarkus, which focuses on nonblocking, fast throughput, and minimal startup time to handle massive concurrent sessions.