DoD thinks it chose one cloud vendor for JEDI. Did it actually?
After months of drama, the U.S. Department of Defense decided last week to choose Microsoft as the solitary cloud vendor for its $10 billion JEDI (Joint Enterprise Defense Infrastructure) project. This despite having previously endorsed a multi-cloud strategy.
At face value, a one-vendor cloud strategy makes sense for DoD: one highly-professional, well-vetted vendor; one cloud; one set of protocols, programs, and drivers; one address for security; one support hotline; etc.
Certainly, history has taught us the pitfalls of a one-vendor strategy. Yet, existing and emerging cloud technology and business models teach us something different: choosing one vendor today may not mean what it used to.
Welcome Back, Vendor Lock-In?
Vendor lock-in has a bad name, and organizations of all sizes – from the private and public sectors - have been fleeing in terror from it for years.
Especially in government contracts, which are always notoriously sticky, stories abound of the evils of vendor lock-in. Take for example the UK Department for Environment, Food and Rural Affairs (DEFRA), which was paying £1.3m per year for over two million individual employee licenses for Oracle - when it actually had only 10,000 employees.
In the cloud too, vendor lock-in has been recognized as poor practice to such an extent that risk mitigation departments in major financial services companies, for example, simply will not allow it. New applications brought online in these organizations must be proven to run in multiple cloud environments before they are approved.
Yet the DoD’s decision to commit to a 10-year project with a single vendor may not be the old-fashioned and potentially wasteful vendor lock-in we’ve known in the past. It’s true that Microsoft is likely to get the lion’s share of cloud business for the near future. Yet given the way the cloud is architected, not to mention emerging multi-cloud technology, this may not be the fixed decision the Pentagon thinks it is.
Technology, not Platforms, Rule the Cloud
In the past, we had to weigh alternatives, adopt a technological platform, then remain loyal to it through thick and thin. We’d choose to work with an Oracle database, and then standardize our infrastructure and applications around this choice. Or we’d choose an IBM mainframe, and adapt our entire business model around it.
This is no longer the case – especially in the cloud. Today, technology no longer defines the platform as technology can be customized to meet our needs. In the cloud – even highly-sensitive, highly-secure clouds for JEDI - storage is essentially commoditized. This leaves the application layer fully open to choosing on a best-of-breed basis.
If I believe that the best serverless compute function for voice recognition is supplied by Amazon, I choose Amazon. If I think that the best AI-driven patent recognition application is supplied by Microsoft, I can use that. I don’t need to swear allegiance to a vendor – just to a technological solution, as long as it provides the best answer to my needs at the time. And if it doesn’t, I can switch.
The commoditization of storage can also create a whole new world of trade opportunities that will keep storage costs down, without compromising on either performance or security. Imagine data storage futures, traded like commodity storage futures, according to pricing and arbitrage between each cloud.
So, although choice of one cloud vendor today by the DoD raises red flags for many about lock-in and the announcement generated a lot of attention, the reality is that a multi-cloud platform will likely be gradually implemented in the long term.
The Bottom Line
Detractors and critics of the DoD decision will claim that by choosing a sole vendor, DoD has hobbled JEDI with technological and commercial disadvantages which will be perpetuated over the next decade.
The fact is that no one – not even the Pentagon – can foresee how cloud will evolve over the next decade. The inherently flexible nature of cloud architecture is what will dictate the evolution of the DoD’s cloud program – not last week’s decision.