On the creeping cost of the cloud

Editor Neil McNaughton compares the 20th Century Telcos’ evolution with the 21st Century birth of the Cloud provider. Both promised cheap bandwidth - comms for the telcos, compute for the clouds. But while comms costs are massively down, compute costs seem to be rising. And the unregulated cloud providers are much more successful at adding bells, whistles and lock-in to their services. With OSDU a case in point. But where is the end user in all of this?

About thirty years ago the big thing was not the cloud but the internet. I attended a meeting organized by what was then France Telcom as it was transmuting from a public utility to ‘Orange’, a private company. The meeting was to thrash-out exactly what should come under the new Orange’s purview. I offered my own suggestion along the lines of, ‘just give us the bandwidth and we will do the rest’. I’m not sure who ‘we’ were at the time. Anyhow my suggestion was very poorly received. There was no way that France Telecom was going to keep a ‘utility’ model in its transformation. It was going to add its own bells and whistles to the internet and it has subsequently done a pretty good job of this. As indeed it has with the provision of low cost bandwidth. A win-win in fact. Or rather a win-win-win as the regulator* has been a guiding hand in a country where fiber connectivity is now available even at the most remote locations and at very reasonable cost.

If the deal in the late 20th Century was the provision of communications bandwidth, the deal in the 21st Century has been the provision of compute bandwidth from the cloud. But the simple value proposition – of vanishingly cheap compute cycles – is much less in evidence as the more or less unregulated cloud providers add their own bells, whistles and more.

A report from Accenture ‘The race to cloud: Reaching the inflection point to long sought value’ fits the paradigm nicely. While the past two years have seen a ‘surge in cloud commitment’, with companies increasing the scope and volume of their cloud initiatives, achieving ‘full cloud value’ is at a ‘tipping point’ with some in ‘cloud transformation limbo’. Only 45% report that they have ‘fully achieved the expected value’, in the form of ‘business enablement’, achieved by ‘unlocking core digital capabilities and ongoing innovation needed to exploit new opportunities’.

This sounds a bit nebulous. But what of the original case for cloud computing, cost savings? Accenture reports that ‘cost savings remain the most elusive of the outcomes’. Only 39% reported fully achieving their cost expectations, ‘perhaps the most frustrating finding for many, as cost effectiveness was one of the early selling points for cloud’. Accenture tempers this finding with the promise of ‘jam tomorrow’. The move to the cloud involves immediate costs but not instant savings. These will come from future tweaks to the business involving modernizing to a ‘cloud-native mindset’, using ‘FinOps’ and a ‘Continuum Control Plane’ to provide ‘transparency’ and manage today’s ‘complex IT environment’. ‘As complexity increases, so too can cost if these critical elements of the value equation are missing’.

While the move to the cloud certainly involves more complexity, I’m not sure that this was very clear up front. Wasn’t the idea to get rid of all that costly in-house IT and benefit from the economies of scale that the cloud providers could offer? Instead we have a labyrinthine stack of technology with a host of experts required to keep the show on the road.

So where does all this leave OSDU, the oil and gas industry’s big push for cloud deployment? I emphasize oil and gas as OSDU started out firmly anchored in the subsurface and, despite subsequent scope creep, its current deliverables are mostly well-focused. That’s OK. What’s not so great is the way the original intent for a ‘a non-competitive API and data store**’ is in danger of being captured by the cloud providers themselves as in the ‘OSDU Data Platform’ on AWS and Microsoft Energy Data Services ‘enterprise-grade, cloud-based OSDU Data Platform’***.

What is perhaps most curious about this state of affairs is that the end user seems to have been somewhat forgotten. If you are a geoscientists sitting in front of a workstation running Petrel, Geolog or DecisionSpace do you actually care about any of this?

* With regard to cloud regulation see our coverage of the ongoing ‘battle’ between Microsoft and the EU elsewhere in this issue.

** Johan Krebbers 2019 interview in Oil IT Journal.

*** There are other offerings, from IBM, from IBM and AWS and a ‘Simple OSDU instance on Google Cloud from EPAM.

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.