Last week at the Moscone Center in San Francisco, CA, Navin Poddar, our CloudBolt West Coast Solutions Engineer, attended Google Next ‘19. He reported back to us from a number of sessions and from the vendor sponsors in booths at the show, a familiar jaunt if you’ve ever been to one of these mega technology conferences.
First things first—networking. Navin ran into his friend, Brian (shown below), who was a Google Cloud speaker. Learning about technology while networking with people could not be better, because seeing the solutions live with people in person and then discussing them informally with peers can uncover more than learning alone.
Anthos and Kubernetes
While Navin was reporting back, some of us viewed the Keynote Addresses to find out what’s new from Google Cloud, particularly after the announcement of Anthos, their new Google Cloud offering that rattled the technology pundits like this one from TechCrunch.
Google CEO, Sundar Pichai, and Google Cloud CEO, Thomas Kurian took to the stage on Day 1 and led with a new message of hybrid and multi-cloud. Whaaaat? Yes, that’s right, these Google Cloud leaders touted their new solution Anthos as all-in on both hybrid and multi-cloud with a signature focus on Kubernetes. They pitched the idea that if your containerization workloads adopt the Kubernetes standard they can be deployed anywhere—including on competitor clouds like AWS or MS Azure. Kubernetes GKE on Prem completes the hybrid cloud story. Thomas debuted as Google Cloud’s new CEO as he introduced his team of product managers who highlighted the details.
The catch, and there’s always a catch, is Kubernetes. You must convert 100% of your virtual machines (VMs) to containers to take advantage of Google’s hybrid and multi-cloud story. But that didn’t stop a lot of us from thinking something slightly different, at least at first.
As Brian Kelly, our CEO stated, “Their multi-cloud message of openness to AWS and Azure made a lot of us pause for a moment; the end result is, of course, to drive more business to Google’s cloud solutions.” He’s also quoted something similar in this News Stack Article.
The reality is that many of our enterprise customers have some interest in considering containerization in production, but they’re not quite there in today’s complex multi-cloud, on-prem, and private cloud infrastructure investments. A lot of production VMs run just fine with multiple applications and services. Containers managed by Kubernetes work best as a one app per container strategy. Converting some VMs could be a roadblock or even a mistake if not done properly. It might be a good excuse to re-architect an application for containerization or best for net new workloads.
As Navin worked his way around the conference, he learned more about serverless computing options for developers as well as about machine learning.
For example, Google Cloud Run, a new offering from Google brings serverless computing to containers. With this option, you run just the “serverless” code that you developed on a fully managed Cloud Run system or you can integrate the serverless code on containers managed by your Google Kubernetes Engine (GKE) cluster. The advantage of the fully managed version is that you don’t have to manage the cluster and you just pay as you go. If you have parts of any legacy application that you want to modernize, you can leverage this serverless approach with containerization that executes it in the cloud.
Anyone hungry? Navin watched an interesting demo of Cloud Auto ML for machine learning. In the demo, the speaker uploaded an image of a pizza to Cloud Auto ML and within seconds it analyzed that it’s 95% likely to be a NY Style pizza and only 1% likely to be a California or Chicago style pizza. These predictions from training high-quality models will soon be the norm to get ahead of the competition for valuable business insights. As with many of the exciting new features announced, this was also in its BETA phase of development.
The industry has zeroed in on Kubernetes as the open source platform to deploy and manage containerization. Google has a bit of a head start on this as it has been using Kubernetes for over a decade. The open source project is now managed by the Cloud Native Computing Foundation.
And BTW, at CloudBolt, we too have been keeping up with Kubernetes. Check out this article from Brian Kelly, our CEO – Kubernetes: Changing the Game for Enterprise Containerization.
CloudBolt has OOTB support for deploying Kubernetes GKE and can manage any Kubernetes cluster.