This article is more than 1 year old

Google puts a gun to the head of IT middlemen – the ops teams

Time to all learn how to code, wrangle AI, we're told

GCP Next If you work in IT operations, Google thinks your role is going the way of the gas station attendant: replaced by do-it-yourself, self-serving tools.

"We want no ops and instead automated IT systems," Google VP Diane Greene told the GCP Next 2016 developer event in San Francisco. "App Engine, BigQuery, our container engine, they just work and developers can be on their way to full 'write and run'."

She cited the example of Snapchat, which went from no customers to 100 million pic snappers without the company hiring an operations team – the IT manager had a couple of helpers and everything else was automated, she said.

This would be a compelling business case for enterprises to switch over to the Google Cloud Platform, Green opined. Google will compete on price, reduce risk from hackers and poor compliance across cloud platforms, and provide access to innovation.

On the cost side of the equation, Google is committed to cutting customer costs as much as possible, and only making buyers pay for what they use. Google's cloud network is the largest in the world, she said, and with more automation, costs will come down further.

As for security, Google is the top hacking target in the world, Greene said, and its security team has developed new tools to fend off intruders. These will all be made available to customers too, and frequently updated.

Finally, on the innovation front, Google is going to make machine intelligence standard across all its cloud services. Alphabet chairman Eric Schmidt bounded on stage to explain how this was going to work.

'Machine learning ... that is the next transformation'

"The platform is not the end of the software process, it's now the bottom layer. Above it is machine learning, with both narrow and general artificial intelligence. That is the next transformation."

Analytics in particular are going to be revolutionized by the growing use of machine learning. Companies will be able to take crowdsourced data, teach computers how to look through it, and reap rewards that a human might have missed.

If you want to build the company of the future, then machine learning and automation are going to have to be a key part of your business plan, Schmidt said. Companies without the technology will be left behind, which is why Google is open sourcing much of its code base for others to use.

As part of this, Google is making a new AI tool available – Cloud Machine Learning – which is an alpha product based around Google's open source TensorFlow deep learning code. This allows developers to set up small AI training programs and test them out before scaling them up to the cloud.

"In the next five years, you will see more change in computing and the cloud than in last decade or two," said Urs Hölzle, Google's VP of technical infrastructure.

"Software is now woven into our lives and we all face the same challenge – how to create apps and operate them at scale cheaply, efficiently, and securely. User expectations are being ratcheted up and apps are constantly being improved."

Managing code operations automatically is going to be a key part of this, Hölzle said, and developers are already having to spend too much time administering code. Google will automate the configuration, provisioning and running of services.

So while the BOFH will probably always be with us, the PFY may have to change his skill set. ®

More about


Send us news

Other stories you might like