Google's Hölzle on why customers shouldn't have to care about cloud infrastructure

Urs Hölzle took to the stage at the 2016 Structure Conference to explain Google's approach to the future of cloud computing.

Image: Jason Hiner/TechRepublic

Google's senior vice president for technical infrastructure, Urs Hölzle, spoke at the Structure Conference in San Francisco on Tuesday about the future of Google's cloud strategy, and why he thinks that customers shouldn't have to care about their cloud infrastructure.

There are a surprising number of Google Cloud customers who have large scale computations where infrastructure is increasingly important, Hölzle said.Things like big data are becoming relevant for almost every industry, which has implications for infrastructure as the scale of data continues to increase.

However, despite the increased pressure that these new technologies place on infrastructure, customers shouldn't have to care about that infrastructure, he said. It should be on the cloud provider to take care of that, so the end user doesn't have to even consider it when they begin new projects.

SEE: Cloud computing policy template (Tech Pro Research)

"You shouldn't have to worry about shouldn't care about this," Hölzle said.

In the ideal situation, Hölzle said, the customer wouldn't need to even think about picking the machine type for their cloud strategy. It is something that Google should determine for the customer--Hölzle called it "infrastructure getting out of the way."

He hopes that, five years from now, only 1% of customer companies know what a machine type is, and 99% feel like they have never had to think about it. The reason being is that the cloud, if effective, should be something that exists almost wholly in the background if it is to properly empower new innovations.

According to Hölzle, most on-premises workloads aren't fully optimized. Moore's law is slowing down, so users may no longer be able to get the same upgrade of power over the same amount of time. This obviously affects hardware, and will likely affect how efficient an on-premises deployment can be.

When a new technology comes around, like Google's TensorFlow Hölzle said, the customer doesn't have to do anything on their end to leverage that technology, if they're using public cloud. That, he said, is the huge upside of using the cloud.

"In the cloud, it's much easier to insert new technology," Hölzle said.

This also includes technologies such as AI and machine learning, which Google has been using for many years, Hölzle said. However, this is contingent, to some degree, on the bottleneck of Flash storage, which Hölzle said is outdated.

The 3 big takeaways for TechRepublic readers

  1. Google's cloud honcho Urs Hölzle said that, if implemented correctly, cloud customers shouldn't have to think or care about the cloud infrastructure they utilize.
  2. In five years, Hölzle said he hopes that only 1% of his customers know what a machine type is, as he feels that Google should be the one determining what is best for their customers.
  3. The cloud also facilitates new technology implementation, he said, as the provider is able to deploy new tech without much work on the customer's end.

More from the 2016 Structure Conference