Goodies from the Google Cloud Next 2019

cl

The cloud computing festivities are on. After Microsoft, it’s Google’s turn, and finally to be followed by AWS re-invent in November. Google cloud next kicked off on April 9. Anyone who is even remotely DevOps would be interested in the event as a great opportunity to understand the new cloud offerings. As a technologist, I am excited about how the latest cloud tech can solve problems in my projects. The competition among cloud vendors is so cut-throat that the pace of innovation is breathtaking.

Here are some of the products I found nifty.

1. Anthos: This was the biggest announcement, and it’s quite impressive both technically and commercially. Anthos promises to power a truly cloud native and cloud agnostic services platform. Anthos will enable developers (enterprise or otherwise) to use a single service running on GCP, to manage their applications on any cloud (AWS, Azure etc.) and even on-prem. Businesses will get a single invoice for all underlying vendors. It is powered by Kubernetes internally.
2. Traffic Director: Google released Istio some time back as a backbone of the latest ‘service mesh’ revolution. Traffic Director claims to be a fully managed network wrapper for service meshes. It will offload all the network heavy lifting like load balancer configs, health-checks, proxies, complex traffic policies etc. from custom services. Teams managing hundreds of microservices will find this tool most useful. Traffic Director will eventually be integrated into Anthos.
3. Fully Managed AI: Organizations are yet to decide whether AI/ML is a fad, but that’s not stopping the cloud vendors from flexing their AI muscle. Google announced a fully managed end to end AI platform that aims to reduce the time from ideation to model testing to deployment of AI models. With Kubeflow, Tensorflow and TPUs, Google wants to create a serverless equivalent for AI (customary vendor lock-in alert). It’s coolest feature is the module that would auto label your data, text, images, videos etc. and create preliminary models that can be iterated upon.
4. Cloud Code: This is something I am personally most excited about. Cloud code at its core is a set of plugins for IDEs like IntelliJ and VS Code to make cloud deployments instantaneous. Developers can deploy directly to Kubernetes clusters right from their IDEs just with a button click, thus reducing deployment times and K8S YAML hell. It will include test and integration pipelines too! Pretty sweet.

There were of course many other APIs and products unveiled at the event, but the above got most of my attention.