Peter Gallagher has a piece in GCN about the reuse of systems components by government. He argues for more incentives for reusable components, saying that when agencies make purchases based only on their requirements or of proprietary COTS packages, they are foregoing benefits to the broader government community. Measuring re-usibility and making it a decision factor in both procurements and business cases are noble goals. Reducing total cost of ownership government wide and increasing reuse has been elusive even after years of Enterprise Architecture efforts and OMB mandates, however. I wonder if a little injection of market forces might do the trick.

Turn now to the emerging trend of Software-as-a-Service. SaaS and its model of multi-tenancy on a common platform is in essence the ultimate in reusibility, and it offers many other TCO benefits. As I have noted in several recent Industry events, however, SaaS is taking off slowly in the Federal government. The business justification of moving capital expenditures to period expenses via SaaS is not applicable due to government accounting rules. Also, government customers, especially in the military, are loathe to have systems maintained and hosted in data centers that they do not control, even when convinced that security is more than adequate (and, as my SaaS vendor friends tell me, often superior!).

As a interim step proposition, I have been introducing the idea of "partnered services" as a way to get SaaS moving in the Federal environment. Based on the industry idea of SaaS as an Appliance (which, I must admit, the same SaaS vendor friends tell me destroys the value of multi-tenancy), partnered service models would put SaaS appliances behind government firewalls for multiple government tenants. This model incorporates several things that I believe will make government customer more comfortable. First, they own the data center. Second, they can acquire it through a procurement model -- likely M&O -- with which they are familiar, and never underestimate the power of making things easier on Contracting Officers these days. Third, they can choose to manage it with a team of vendors with which they are familiar. Imagine a team of the "resident" IT services vendor deeply ingrained in the agency mission and legacy systems partnered with the SaaS vendor, with incentives to deliver the service to agency customers. This model may initially remove some of the benefits of SaaS, but it is a necessary first step in the ultimate migration.

Best of all, once these SaaS services are in place, what's to stop an agency from offering it to other agencies? Rather than building their own, agencies can choose to buy services from each other. The providing agencies will work to make their services the best since they recoup budget dollars via selling them. The resulting market forces will increase standardization and re-use and reduce total costs to the government.

DISA is piloting a similar model in the NCES collaboration product line, where they are offering two competing services to users. DISA's vision of providing standardized, reuable services is a good example of an environment where partnered services could grow.