You may already be familiar with my ongoing frustration with Andrea DiMaio, and his relentless effort to ensure that government IT staff around the world is afraid of failure, change, and the future. This pessimism, especially coming from Gartner, contributes to the culture of conservatism and risk-aversion that confounds the same reforms he claims to support. In his most recent note, the fear starts at the headline: Why Government IT Consolidation May Be Doomed.
Yikes. It’s a wonder I why bother waking up in the morning.
His thesis is this: Since data center consolidation efforts and cloud services are both handled by “shared services” offices, those offices are at cross-purposes and may fail as a result. They will lose money as their customers will find clouds more attractive, robbing them of the budget they need to perform the consolidation work. Thus, government IT consolidation may be doomed.
I think the story is a lot more hopeful than that.
There is no conflict.
The first assumption is that the shared services organizations are also handling the cloud services contracts. This is rarely true. In the US government, at least, cloud contracts are mostly negotiated at the agency level or consolidated under massive, government-wide blanket purchase agreements like the GSA IaaS contracts. Shared services providers live further down the chain, at the program level.
But even if a shared services provider was delivering a cloud contract as well, it’s not a problem.
Cloud is consolidation.
Cultural preferences, procurement rules and security concerns go a long way to ensure that agency data will stay inside that agency, drastically reducing the number of possible competitors for a program’s business. Even if I wanted to move my entitlement eligibility system to Amazon, I can’t. Instead, I’ll demand Amazon-like services from my colleagues in Shared Services.
In response, the shared services providers are using the consolidation mandates as an opportunity to make themselves more cloud-like. Some, like NASA Nebula, go so far as to implement the Amazon EC2 interface. Others, like DISA RACE, are content to virtualize and automate some of the more labor-intensive aspects of the data center and call it a day. In either case, providing cloud services is part and parcel of their consolidation effort. The notion that the two efforts are opposed is misguided. Most of the shared services shops I talk to want to provide their customers with both on- and off-premise infrastructures, and make it easy for those customers to move between them.
Also, budgets don’t work like that.
DiMaio’s argument relies on shared service providers having a fixed budget to deliver their services. This is almost never the case. Organizations like the Defense Information Systems Agency or the National Business Center at the Department of the Interior rely on a variety of funding vehicles. Some may make an investment up-front, but have to recover that investment within two years. Others will charge the cost of the service plus a prescribed markup for overhead. Others will be provide little more than space on a data center floor, which is a fixed cost they bear anyway, and charge rent. The most sophisticated will fund large infrastructure projects by pooling the infrastructure budgets from their customers. There are all kinds of ways these providers are funded, and they often employ many strategies at once. Introducing cloud computing services doesn’t mean ruin, it just means smaller data centers — which is the point of the exercise. These shared services funding models are more flexible and scalable than he assumes.
Less fear, more innovation.
So this is the story: faced with breakneck consolidation schedules and really disruptive technology in the private sector, these shared services organizations and their respective CIOs are doing everything they can to make the change worthwhile. We’re seeing organizations rationalize their application portfolio, shed legacy processes and the software they demanded, and give themselves a modern, consolidated, shared infrastructure that can take advantage of public cloud services to become drastically more manageable, flexible, and scalable.
Data center consolidations are a nasty business. There are winners and losers, and many efforts will fail. That’s just how things work. But rest easy, government IT staff. You don’t have to dwell on that, because DiMaio is wrong. You a long way from being doomed. Instead, you have a remarkable opportunity to change the way that we’ll manage government IT for the foreseeable future.