Hurry up, Google -- we need a strong third cloud

With machine learning, Spark analytics, container management, and big global expansion plans, Google Cloud may soon nip at the heels of AWS and Azure

Hurry up, Google -- we need a strong third cloud

Last week's conference for Google Cloud Platform users, GCP Next 2016, spoke to Google’s relative standing in the cloud competition: Roughly 2,000 people attended, as opposed to the 20,000 crowding Amazon’s last AWS re:Invent show, and many of Google’s announcements played catch-up rather than led.

Yet that trailing position may be temporary. GCP Next left no question that Google is serious about the enterprise cloud -- "dead serious," in the words of Diane Greene, Google’s top cloud exec. Hired just last November, the celebrated founder of VMware laid it on the line: “We've spent billions on data centers and are going to use them as much as we can. This is a long-term, forever event," proclaimed Greene.

Infrastructure buildouts on this scale speak volumes in the cloud world. The precursor to Microsoft Azure's emergence as a cloud to be reckoned with, for example, was a massive investment in cloud infrastructure. At GCP Next, Google announced two new cloud regions: a U.S. Western Region in Oregon and an East Asia Region in Japan. That won’t bring GCP up to snuff with the regional coverage of AWS or Azure, but the 10 additional regions planned by the end of 2017 should do the trick.

The juiciest announcement at GCP Next was the new Cloud Machine Learning service, based on Google’s open source Tensor framework. Alphabet chairman Eric Schmidt waxed eloquent about the impact of cloud-based machine learning on the future of the technology industry, though he neglected to mention machine learning services have been available on AWS for months and on Azure for over a year. (We’ve reviewed both and will evaluate Google’s offering soon.)

Google also announced the general availability of Cloud Dataproc, a managed Hadoop and Spark platform first announced last September. Again, Google is following in the footsteps of similar services offered by AWS and Azure, but Dataproc adds heft to an already extensive line of data-intensive cloud services offered by GCP, including BigQuery, Cloud Dataflow, Cloud Datastore, Cloud SQL, and Cloud Bigtable.

Perhaps most important of all in Google’s enterprise push is the rapid evolution of Kubernetes, Google’s open source container management project. Kubernetes 1.2, released earlier this month, offers a new GUI front end along with four times the scale of the previous version. Kubernetes could provide a hybrid on-ramp for enterprise customers: Run container-based workloads on Kubernetes locally and lob them as needed to Google Container Engine, GCP's Kubernetes service in the cloud.

You can view Kubernates as part of what Sasha Labourey of CloudBees calls the "decomposition" of PaaS (platform as a service) into standards-based Docker containers, open source container managers such as Kubernetes, and open source dev workflow engines such as Jenkins. Indeed, Eric Schmidt went so far as to admit that Google's AppEngine PaaS was a mistake, because it forced developers to both build and deploy Google's way, on Google's cloud.

Containers and their management solutions are all about portability: Build it, package it, and run it anywhere. Interestingly, Google and Red Hat made a joint announcement in January that OpenShift Dedicated, a cloud version of Red Hat's OpenShift PaaS, will soon be available on GCP (it's already available on AWS). Not only is OpenShift sufficiently "decomposed" that it supports both Docker and Kubernetes, but Red Hat is also the No. 2 contributor to the Kubernetes project behind Google. Open source Red Hat software running on-premises in concert with GCP and Google Container Engine in the cloud could be a powerful hybrid combination.

Other signs of enterprise cloud seriousness include a new Cloud Identity and Access system, an audit logging solution available in late May, customer-supplied encryption keys, and the StackDriver monitoring, logging, and diagnostics service (which works with AWS as well). Google also announced some new GCP customers, including Disney, Home Depot, and Coca Cola.

Much progress has already been made with GCP, but much remains to be done. The question with such a long to-do list is always: Can they execute? Outside of its search and advertising business, Google has often seemed all over the place, spinning up quirky projects and pulling the plug on others that people relied on. But Google is also in a virtual tie with Apple as the largest company in the world. When it sets its hive mind to something, it has the brilliant engineers and colossal resources to back it up.

Copyright © 2016 IDG Communications, Inc.

How to choose a low-code development platform