Talking cloud

Posted by on July 2, 2014 4:53pm

Five ways cloud computing is going from red hot to white hot (part one in a series)


I’m just back from GigaOm’s Structure 2014 , my yearly appointment with the cloud industry. GigaOm managed once again to gather a formidable lineup of thought leaders, movers and shakers, providing an expedited way to get the pulse of the industry and connect with the cloud tribe.


While everyone at this event is focused on the cloud, this year they were also focused on the crowd. This year’s conference was standing-room-only, a clear indication that cloud technology is getting more mainstream. Every company today needs to have a cloud strategy. Even companies slow in adopting the cloud, certainly can’t ignore it any longer.

As I reflect on the event, I see five key trends driving this sector over the next 12 months. For now, I’ll highlight one of the leading points of discussion, the trend of automated IT management.

Automated IT management essentially takes the traditional work of IT and system administration — configuration, maintenance, updates, new allocation, provisioning, fixing, optimization and so on — and has software complete the tasks. It means leaving to software all the educated decisions that must be made in the course of running an IT system. Imagine collecting all data available from IT systems today and feeding them into adaptive systems that have the capability to make intelligent decisions. With the availability of big data and machine learning algorithms we have today, it’s not unconceivable that all of those optimization decisions could be made on the fly by highly sophisticated technology.

We know, of course, that virtually all industries have been disrupted by software and the cloud paradigm; now we’re looking at IT itself being disrupted entirely.

It’s highly plausible. As Khosla Ventures founder Vinod Khosla noted at the event, “It’s ridiculous to have humans manage the level of complexity that they do. Let’s take that out”.

Now, consider that the largest IT cost today is cost of ownership. People often focus on hardware and software costs but these are small in comparison to IT personnel costs. By bringing machine learning and data analysis techniques to IT itself, software could be making the large majority of IT decisions and humans can be removed from the IT equation entirely. IT costs would fall almost exponentially.

The advantage of this new paradigm of automation is not only its cost savings opportunity, but most importantly the way it would enable ‘adaptive IT,’ As the rate of change for businesses accelerates, companies are less able to predict where business is heading. The top strategic imperative becomes responding rapidly via agility and modularity through “software defined-everything” and adaptive IT. When you don’t know what the future holds, it’s important to have the infrastructure in place that can respond to business needs on the fly.

Diane Bryant, SVP and GM of Intel’s Datacenter Group, noted that in a true software defined infrastructure (SDI), the point of optimization is no longer at the server level, it’s at the rack and the cluster level, for a composable and configurable on-the-fly infrastructure including I/O, storage and compute.

And Urs Holzle, senior vice president for technical infrastructure at Google, pointed out that this next level of automation just the latest impact of Moore’s law, and will help remove the constraints currently imposed by the cost of labor.

‘Moore’s law is relentless,’ he said. ‘Unfortunately, it doesn’t apply to IT budgets. IT has to continue to do more with less and the only way is towards automated IT management.’

I’ll share some of my other key takeaways from Structure in future posts.

Don’t hesitate to contact me directly with your comments and inputs via paola dot moretto at nouvola dot com. You can find me on Twitter at @paolamoretto3 or @nouvolatech.