Happy New Year. We wont be on twitter much, except when something big happens or during re:Invent 2018. We wanted to leave you with few thoughts.
- Serverless is the future. Accept it.
It was not a popular thing to say Cloud is the future back in the days. If you thought Cloud was inevitable, the drivers behind serverless are even more compelling. Migration to Cloud from DC was a harder choice than migrating from a VM based app to serverless app. Developers will start experimenting with things like Lambda and will end up liking the potential of it. The amazing Amazon machine will continue pushing it and it won’t be long we will see more and more workloads migrate to AWS Lambda. My hope is that Microsoft and Google will wake up from their deep sleep and start increasing their investment in their own serverless offerings sooner than later. Because it would be a terrible thing again to give the entire field to AWS and watch on the sidelines for several years.
- Containers for new workloads is an anti pattern
Using containers today is like using CloudStack or OpenStack in 2012. Exciting, but limited future potential. If you have an existing workload that can benefit from containers, go for it as a short term stop gap measure. Containers can not be your long term tech strategy. The faster you move away from VM based apps, the better positioned you will be for the future. Your goal should be to reduce the number of things that your ops team has to do or manage that are at a low level.
- Do not fear vendor lock-in, architecture lock-in is worse
Vendor lock-in is not the hardest thing to overcome, Architectural lock-in is harder to overcome. If you built your new app components today optimizing for constraints of a VM, you will have a harder time moving to future than migrating an app from AWS to GCP. For example, using Kubernetes for new workloads creates an architecture lock-in that you will have a harder time getting out of it and move to serverless. Even people migrating off of Oracle tech have reaped plenty of benefits from using Oracle stack for last 10–15 years. The current benefits of committing to a platform outweigh the future cost.
- Pick a platform and go full onboard
There are 3 platforms that matter for the immediate future. These are AWS, Azure and Google Cloud. Pick the one suitable for your company workloads and stick with it for next 5–7 years. There is no wrong answer here. Even with Google, you can go serverless by leveraging several things like Google App Engine and the eternal beta Cloud Functions. On Azure, you can use Azure Cloud Functions. You may not be on the cutting edge as using AWS Lambda, but in the longer horizon, it won’t matter. They all innovate when market signals emerge and bridge parity in due time. Pick one of the three and stick with it. ( we like GCP currently because of their customer friendly pricing model).
Thats all folks.