Google Cloud Weekly 8/2/2018 – Info Linux

Google Next 18 is officially in the books, and there were a ton of new announcements, products, and features revealed to enhance the functionality and value of Google Cloud.

Without a doubt, the week’s most exciting announcements were centered around two of Google’s strengths. Google Kubernetes Engine and Machine Learning.

Starting with Kubernetes Engine, Google announced the alpha release of Cloud Services Platform. Cloud Services Platform is a managed service that runs on top of Istio, which is an open source framework used to manage container microservices across multiple platforms. Cloud Services Platform lets you manage both GKE clusters and on-premises Kubernetes clusters from the same interface.

A common theme that we saw throughout the conference was Google’s ongoing commitment to open source frameworks underpinning many of their products. Their belief is that open source makes you (1) futureproof, and (2) able to work with a greater variety of tools.

Related to Cloud Services Platform, Google also took things one step further by announcing GKE On-Prem. With GKE On-Prem, you can deploy a Google-configured version of Kubernetes, both on-premises and in other clouds, which lays the foundations for true hybrid computing. Think of it as turning your on-premises Kubernetes environment into it’s own version of GKE, even though it’s not (yet) in Google Cloud.

But wait, there’s more! Google also announced a new GKE serverless add-on built on top of Knative, another open source framework. Noticing a trend here? GKE serverless add-on lets you run serverless workloads, such as those that respond to events, on top of Kubernetes Engine. So if you’re stuck on whether you want to work with serverless or Kubernetes, now you don’t have to decide. You can use both.

Moving on to Machine Learning and AI… We saw the expansion of AutoML to cover not only Vision for image recognition, but also Natural Language and Translation. In case you’re not familiar, AutoML lets you apply custom machine learning to suit very specific needs, without having to code your own machine learning model. This means that you don’t have to be a data scientist, or be an expert in machine learning programming to create custom models.

We also saw the exciting announcement of BigQuery ML, which lets you run machine learning models directly against BigQuery datasets, using simple SQL commands. That’s right, run machine learning models directly in BigQuery, and with the insane performance that BigQuery offers. Previously, you had to export your BigQuery data for machine learning analysis. This will be a big improvement.

Finally, we saw the release of Edge TPUs for the internet of things, which is a purpose built machine learning chip that runs a light version of Tensorflow directly on your IoT sensors for accelerated machine learning. Developer kits will be available in the near future.

There were, of course, many other announcements at Google Next 18. If you would like to see the entire list of new products and features, check out the link here for more info. Thanks for reading, and we’ll see you next time for Google Cloud Weekly!