While in Portland for OSCON a few months ago, I was invited to take part in an open discussion with a couple of colleagues, Krishnan Subramanian and Alex Williams. The idea of the session was to have a free ranging discussion about the future of the cloud – the three of us don’t often have the chance to engage face to face so it was great to sit down and chat.
Much of the conversation centered around the notion of federated clouds, and how much of an important part open source can play in that. In fact this part of the discussion started a separate conversation in which I suggested that;
history has shown that commercial imperatives win, and there will always be some kind of block to completely free federability.
Krish wrote a well considered response where he countered my argument. Bottom line is we both agree open source is lowering the barriers to entry but time will tell the relative success of open versus closed approaches towards the cloud.
Outside of the open source debate, there was a lot of discussion about need to move away from simply looking at cloud computing as a way to save money. I managed to bring out Jevon’s Paradox and use it to justify my position that organizations should not be looking at the cost implications of cloud, but rather looking to how cloud enables;
- Much more rapid innovation, experimentation and the general democratization of technology
- The ability to closely tie expenditure to revenue
We spent a bit of time talking about what organizations track when utilizing the cloud. I reflected on the fact that both the best and worse things about cloud are the fact that it enables business units to deploy there own services. Good because it far reduces time to benefit but bad because it means controls can be bypassed. This is a real barrier to adoption and why IT is spending time blocking cloud adoption. We need to find ways to deliver insights in terms of cloud expenditure and use, but also to tie that usage to bottom line revenues and benefits.
Alex and Krish took a well-aimed swipe at IBM and their approach of bundling big iron with software and calling it a cloud. I countered by saying that cloud is all about allowing people to use technology in a subscription based manner – I believe private cloud enables organizations to internally offer their business units the ability to rapidly provision technology and have some internal cost recovery apportioned.
Finally we talked a little about PaaS and where it currently is in its lifecycle – I think we all agreed that we we are very early on in the growth of PaaS.
It was a fun session, the building it was held in (Wieden + Kennedy’s Portland offices) was amazing and the drinks afterwards were awesome – check out the video below.