Trends in Cloud based engineering

The move to "cloud" has brought in new innovations in software development and deployment. Here are some of my observations on where things are headed in cloud-based engineering space.

Hybrid Cloud

Public cloud is everywhere. This is old news. Hybrid cloud is real, big enterprises are embracing it and so are cloud providers. In the past years, public cloud providers have realized that the largest revenue sources are large enterprises who cannot just move (even if they wanted to move). AWS Outposts, Google Cloud Anthos, Microsoft Azure Stack and Vmware Cloud are tackling the hybrid cloud market with gusto. Their mass adoption strategies are different. AWS is pushing hard to get customers on-boarded to proprietary solutions, whereas Google is co-building the technology with the industry with hopes that this will lead to sustained growth and a longer lasting ecosystem. Microsoft Azure is busy leveraging Microsoft’s huge customer base to provide proprietary solutions in this space. Vmware also has a big presence in enterprises and it is selling hybrid cloud solutions to keep its customers tied into its ecosystem. In some sense, the relationship of Vmware with Cloud vendors is both symbiotic and competitive.

Microservices

From large monoliths to Service oriented architecture (SOA) to Microservices - software engineering methodologies are changing rapidly. Microservices improve business agility and developer productivity but bring in a whole new class of problems - discoverability, distributed debugging, security, end-to-end SLAs, deployment & rollback, etc. Fortunately for those problems, there is a whole class of open-source and proprietary solutions being built. Given the momentum, I think in the next 3 years, we will have a good (integrated) ecosystem around Microservices based software engineering. Right now the DevOps effort around Microservices is non-trivial.

Developer Experience/Productivity

Every large company I’ve talked to has an organization dedicated to developer productivity. The organization serves various needs from language support to build to CI/CD to software release infrastructure to common tools. This area is fractured due to the large number of (non-integrated) options for doing various things. And public cloud(s) complicate it further with their own tools. The current tool integrations are clunky, hard to use, brittle and inflexible. I think this area is ripe for innovation for end-to-end drag-and-drop solutions. There will always be a need for a Developer productivity team, and as software stalwarts we should try to cap it at a team, rather than an org.