Earlier today, chatting with a European Gartner sales colleague who was telling me how cost containment has become almost the only topic for discussion with his government clients, he told me that the jurisdiction he is responsible for is in discussion with other jurisdictions to develop some form of cross-boundary community cloud infrastructure (see here for the NIST definition of community cloud).
This looks like yet another example of “government cloud”, that is a community cloud restricted to government organizations, although its cross-boundary nature may make it a quite intriguing example, given that most of those I have heard about so far are limited to a single country.
However, in relating the emergence of government clouds to the original topic of our conversation (cost containment), it occurred to me that building or sourcing a community cloud will inevitably take some time, and the number of client organizations and how extensively and how fast they will join such cloud is not entirely predictable. Most of the savings and the other advantages of cloud-based infrastructures – such as flexibility and elasticity – depend on their scale. The stories we hear about savings with Google or Salesforce.com or Amazon are not the same we’ll experience with government clouds, because of their smaller scale and higher levels of security.
So, if the primary driver for considering cloud-based solutions is cost, one may argue that helping each agency segment its requirements (in terms of low, medium or high levels of security, and corresponding levels of control) and encourage them to adopt public cloud approaches for the former could yield greater savings that moving everybody onto a government cloud.
The counterargument is that the government cloud provider (or providers) should offer different service levels for different workloads, and source low-security ones to public cloud providers. While it is quite likely that many government clouds will be hybrid, i.e. using a combination of own virtualized infrastructure and external providers, it will take time before they can establish themselves and articulate both a clear offering and a cost-effective sourcing strategy.
The challenge for governments worldwide is – once more – to find the right trade off between a strong mandate to centralize and a common framework for each agency to source its own infrastructure and application requirements. The government cloud should be one of the cloud sourcing options, but not the only one.
Category: cloud Tags: government cloud

Andrea Di Maio





































































































2 responses so far ↓
1 Tweets that mention How Big Should Government Clouds Be? -- Topsy.com June 30, 2010 at 6:22 am
[...] This post was mentioned on Twitter by govwiki, Fabasoft eGov, Toshio Matsuda and others. Toshio Matsuda said: [Gartner] How Big Should Government Clouds Be?: Earlier today, chatting with a European Gartner sales colleague wh… http://bit.ly/apLDIO [...]
2 CloudWatch.dk » Læser Gartner danske arbejdsprogrammer? June 30, 2010 at 3:44 pm
[...] et – som sædvanligt – oplyst og læsevenligt indlæg på Gartners blogsite, skriver Andrea Di Maio om et land, han har hørt om, der er i færd med at undersøge, om det [...]