Current cloud application platforms for enterprise software are like “radio on TV”.
Like early TV programs that showed a (radio) announcer reading news off a page, the current aPaaS take the good-old models of application architecture and programming (basic request-driven SOA, Java EE, .NET) and place them into the cloud. The PaaS frameworks (CloudFoundry, OpenShift) provide some cloudiness, but the fundamental architecture of applications and application containers does not change. (Richard Watson provides excellent advice on how to create well-fit cloudy applications in the constraints of the older platform architecture)
Enter AWS Lambda. An application platform that is event-driven, microservices-based, firmly decoupled from client considerations, elastically scalable by its core design, real-time and cloud-natively priced. A cloud-native vendor introducing a cloud-native platform architecture.
That this is a significant innovation can be witnessed from just the fact that shortly the mega-vendors with enterprise ambitions announced their own renditions, including Google Cloud Functions and IBM OpenWhisk. Expect most of Oracle, Microsoft, Red Hat, SAP and others to follow soon.
I believe, we are witnessing the beginnings of the first true born-on-the-cloud platform architecture. The offerings are still immature and the actual sustainable implementations will take time to arrive. Amazon’s Lambda is limited in function and scale, IBM’s OpenWhisk is a research preview, others — still in the labs.
Yet users and IT planners should pay attention.
In 1997, we published a note entitled “Greater Java: a Continent Emerging?” that anticipated the emergence of Java EE at the intersection of CORBA and the then just introduced Java. I believe it’s time for the new note entitled “Born-on-the-Cloud Microevent Application Platforms: a Continent Emerging”.