For governments to realize the benefits, however, they must take much of the work they have been doing for themselves and outsource it. True, this outsourcing may simply involve consolidation across the local, state or federal enterprise, keeping things within the government tent. But some of it will need to go to private-sector providers that can serve communities larger than any given government.
The transition to outside providers naturally makes governments nervous about the dangers of dependency. What if the outside providers don't handle security or privacy in ways the public expects and demands? What if, over time, the outside providers are quite efficient but lock in their government customers to rates that unfairly grab too many of the benefits for themselves?
Such concerns have resulted in a slower takeoff into the cloud for governments than many were predicting a few years ago. Still, cloud penetration will continue to grow. Ten years from now, more than half of the information-technology jobs within today's government enterprises will have shifted to larger-scale cloud providers.
While all this is true, it conceals a critical problem we are presently missing, not only in the United States but also in other governments, such as those of Canada and Australia, where the layers of government are substantially independent. The problem is that the national, regional and local entities are working on their own cloud applications and, by and large, ignoring the benefits of working together. As a result, we're much slower than we ought to be in mobilizing movement to the cloud. This is particularly true of local governments, which, though smaller individually, are huge in their needs for better IT services. In addition, we're headed to a future that keeps government demand fragmented in ways that advantage cloud producers but disadvantage cloud consumers, including the public.
Fortunately, remedies are straightforward and have been applied in other settings. Fundamentally, we need the three levels of government to work in the following way:
• Local governments need to move to the cloud while simultaneously collaborating. Local governments are under the greatest pressure for productivity improvement and cost savings, and they will move if they can get seed money and support to get over the transition hurdle. Localities that used to bitterly oppose outside influence are now adjusting to a new and more interconnected world. Independent school districts that can no longer afford to offer the variety of courses they used to or that their students and parents now demand, for instance, increasingly are offering them through cloud-delivered courses. (It's estimated that by 2020 half of all K-12 courses will be provided via the cloud.) And local governments need to standardize their demand and aggregate their buying power, whether they are merging IT services or hosting them for other jurisdictions, as Oakland County, Mich., is doing.
• Regional and state governments need to extend cloud-based production to serve local jurisdictions in their realms. This can apply to networks and applications at the lower end of the technology stack, as with states getting dispersed facilities connected by broadband to serve as "anchor tenants" for broadband access to the surrounding communities. This also applies to cloud-based educational applications and to GIS, licensing and other applications that localities can't afford on their own (Michigan is collaborating with local governments on a number of these possibilities). Expanded state production benefits all jurisdictions, raising the effort to the scale required for efficiency in a cloud-based world.
• National governments need to provide seed money and experiments with a view to standardizing at the scale required for the cloud. This, of course, is exactly what happened when the Defense Advanced Research Projects Agency funded experiments leading to the initial packet-switching TCP/IP standards upon which the Internet is built. These days, much of that kind of research at the federal level is focusing on cybersecurity in the hope that advances will ripple out to other levels of government and to the private sector. These R&D efforts are best suited for the national level: Standardization for government cloud-based production and consumption is critical, and it needs a greater priority than it is getting. It wouldn't cost much, and could make all the difference over the next five to 10 years.
Many government problems have resisted IT-based innovation too effectively and too long. Cloud computing holds a lot of promise for changing that. It would be a shame, however, if excessively incremental thinking doomed us to miss the best cloud-based opportunities, the ones that require us to work more closely and effectively together.
