There have been some interesting debates on cloud computing over the last few days. I thought I'd mention Tim O'Reilly's and Hugh MacLeod's posts.
I left a comment and I've summarised my thoughts here just in case my comments don't make it onto their blog posts (this happens quite often).
Comment =====
The question of large scale monopolies in what is today called "cloud computing" has been raised many times before in the past by Tim O'Reilly and others. There does exist a real danger here of network effects particularly through mechanisms such as aggregated services (i.e. market reports, preferential pricing or performance) along with the more obvious danger from the creation of a large proprietary technology based cloud.
None of this has been kept secret, it however has only just recently (in the last few years) become more popular. My reasoning, back in 2006, for the need to open source Zimki (a now defunct utility computing cloud) was based upon much earlier conversations that were raging about the dangers of a lack of portability & interoperability in a future utility computing world. None of this was new stuff then, it was just not widely known.
The move towards the cloud is almost inevitable (the business equivalent of the Red Queen Effect) due to cost efficiencies and speed of new service release. The use of open sourced standards (operational open source reference models of services to be provided) such as Eucalyptus (as Tim points out) is a way to prevent the formation of artificial markets, though as JP notes it does require concerted community action. Such action is in the interest of general business consumers for all the reasons of second sourcing.
As for the connection between cloud and services (as in service oriented design, architecture and "as a service") this is simply the shift of a number of business activities along the S-Curve of transformation from innovation to commoditisation. This is also allowing for greater componentisation and hence new innovations such as mashups etc. It is all connected by the same underling process and a tri-partite effect of participation, network effects and componentisation which is both accelerating innovation in the IT field and commoditising that which is ubiquitous. The pipeline of creative destruction has accelerated and this is what connects it all.
Finally, James Governor's comment that "Customers always vote with their feet, and they tend vote for something somewhat proprietary" whilst insightful is also somewhat unfair. It was clear from CloudCamp London that potential consumers are concerned about the issues of interoperability and portability and none thought that proprietary technology would resolve this issue. I feel compelled to defend those users, as the assertion that James puts forward is that the average CIO lacks the strategic wit of their counterparts in manufacturing.
This might be true, James might be right and maybe CIOs of today have become lazy on the easy pickings of IT. In such a case, the move by some companies over the last few years to put former heads of manufacturing in charge of IT and to bring in those skills honed through years of cut-throat commoditised activities and the strategic imperative of second sourcing may well turn out to be wise moves indeed. It would be unwise however to simply focus on the commoditisation of IT as both extremes of innovation and commoditisation need to managed simultaneously.
I remain hopeful, that those governing consumption of IT within business will realise the need to form consumer associations to push the industry towards open sourced standards for ubiquitous activities. What I cheekily called "Gang up now before the *aaS cloud gets you"
Time will tell.