Google announced a strategic alliance with Intel that Google aims to rely on to make it more competitive in the enterprise cloud market.
Enterprise cloud is still the smaller part of the overall cloud business, but it's also growing at a very healthy clip -- over 20% CAGR through 2020 by one estimate. Amazon Web Services Inc. 's ownership of roughly a third of the public cloud business seems unassailable in the short term, which makes enterprise cloud an attractive target for all the cloud also-rans, the most prominent of which are Google (Nasdaq: GOOG), IBM Corp. (NYSE: IBM), Microsoft Corp. (Nasdaq: MSFT) and Salesforce.com Inc.
IBM, Microsoft and Salesforce are all coming at cloud from the enterprise side. Google needs to simply keep up with its rivals, but with the growth that enterprise cloud (and hybrid enterprise-public cloud) is experiencing, Google has a legitimate expectation of earning some extra market share if it can improve its enterprise cloud products.
The alliance with Intel Corp. (Nasdaq: INTC) follows hard on the heels of Google's announcement on Tuesday of its formation of an artificial intelligence (AI) research group within its Google Cloud operation. Google had to share headlines with Microsoft and Amazon, which also had cloud AI announcements. (See Hypercloud Guys Make Big AI Push.)
Improvements in the electronics industry are almost always predicated at least in part on advances in computation performance, and that's where Intel comes in.
here on Light Reading.
The two already collaborate on data center infrastructure, but according to the blog post announcing the collaboration, Nan Boden, Google's head of global technology partners, said the two "are now expanding our collaboration to help enterprise customers move from legacy infrastructure to an open, secure and future-proof cloud."
The two companies defined several areas where they will collaborate. The first, they said, is on optimizing Kubernetes for Intel architecture to support a broad range of workloads. Kubernetes is the open source program Google created in 2014 to develop application containers. Intel is a major contributor to the Kubernetes ecosystem and enables enterprises to run OpenStack as a managed Kubernetes workload. The goal of the allies is to improve workload capabilities, including virtual network performance and prioritization of shared resources.
Google and Intel plan to improve TensorFlow performance on Intel processors. TensorFlow is a means of allocating elements of computation to different processors in an array of processors, whether within a server farm or a single device, such as a laptop or a smartphone. Google provided an example in which the technique could be applied within a single multicore chip. TensorFlow is an outgrowth of Google's Machine Intelligence operation, but Google says it has applicability beyond machine learning. TensorFlow too is now an open source program.
The two will also address security issues in computing in general and in Internet of Things (IoT) applications specifically. They said they intend to create a secure platform for connecting Intel's IoT edge devices to Google Cloud Platform (GCP). Google and Intel also said they would enhance security integrations between Intel hardware and GCP infrastructure that will further improve security for enterprise customers.
As for Intel, it's nearly impossible for that company to burrow deeper into the data center server market, but it has a near-monopoly to protect. The computing world is going open source, and that seems to play against Intel's strengths, but its collaboration with Google could be a savvy counter-maneuver. Intel is looking to make improvements in its processors explicitly to support open source efforts such as Kubernetes and TensorFlow.
— Brian Santo, Senior Editor, Components, T&M, Light Reading