Over the last week I have speaking with several vendors who are developing or restructuring their “cloud” service offerings for government. What I have noticed is the increasing popularity of the term “private cloud” as in “we’ll help you build your own cloud on premises” as opposed to “we will run a private cloud for you off premises”. Actually earlier this month I read again about the US Army’s email migration into the cloud. According to NextGov
The Army will begin migrating to an enterprise e-mail service in the middle of February with 1.4 million unclassified and 200,000 classified accounts moved by the end of the year, according to a report by the Army News Service.
According to Government Computer News, predicted savings are in the order of $100 million per year, taking into account that the current cost per use will drop from $100 to $39 per year, and users will have better functionality, ranging from an increase in mailbox size to the ability to access their email anytime anywhere.
From the news I have read (I have not had a chance yet to speak to anybody at the US Army or at DISA), this looks like a sensible migration, where there is no change of email system (they will remain on Exchange) but significant improvements in service levels and value for money.
However, this is yet another instance of a private cloud. Of course this will leverage the excellent work that DISA has been doing for a long time with its RACE infrastructure. But will they run a private cloud only for the Army, or will their infrastructure also support email for other agencies or forces going forward, hence become a true multi-tenant community cloud?
It seems to me that the migration to a private cloud is “relatively” painless with respect to more traditional types of infrastructure or application migrations, although it requires a new way to consume and possibly pay for services. If your provider is either your own IT department or another government agency running a private cloud for you, many of the stumbling blocks about security and control can be overcome.
But all this leads to an interesting question.
Once many government agencies will run (or have somebody run) their own private cloud, what incentive will they have to move to community or public clouds that consolidate and replace those? Isn’t there a risk that what is being done today will soon become a legacy of private clouds that won’t allow government as a whole to realize but a fraction of potential benefits?