After fifteen months of proposals, protests, and political wrangling, the massive DOD JEDI cloud services contract was awarded to Microsoft on October 31. While the inevitable protests will likely delay the final nail in the coffin, we are clearly moving out of the contracting phase and into the implementation phase.
Many DOD agencies are probably asking: Now what?
While the operational requirements of the Joint Enterprise Defense Infrastructure (JEDI) cloud offering are laid out in some detail in the JEDI RFP, the process of actually implementing cloud migrations across large swaths of DOD is far less clear. Most Department of Defense agencies are not prepared – culturally or technologically – to take full advantage of the speed and agility of a commercial cloud offering.
At BlueCat, we’ve guided plenty of enterprises through their cloud journeys. From our perspective on the network infrastructure side of things, we’ve seen constructive approaches and…well…not so constructive approaches. As DOD starts to wrap its head around the next phase of the JEDI cloud, we thought we’d offer some thoughts on how to maximize the value of this brand-new asset.
Change DOD’s cloud culture
After years of caution and deflection in discussions about the cloud, DOD agencies are going to have to get up to speed quickly. This isn’t as easy as it sounds. While most network administrators in the military recognize the intrinsic value of the cloud for commercial companies, it will be a mental shift to make cloud the default choice for government networks.
The detailed treatment of security, availability, and functionality in the JEDI contract was designed to overcome those concerns. Administrators now have the top cover to make cloud conversations possible.
Yet as any DOD technology provider knows, old habits are hard to break. Moving from the knee-jerk stance of “we don’t do cloud” to embracing a DevOps culture will be a huge leap. Some agencies have experience with MilCloud or have worked with cloud-native units like Kessel Run, but they are in the minority. Most will have to go from zero to cloud in far less time than they’d prefer.
The inevitable reorganizations and realignments are coming, but these won’t create a new cloud culture in isolation. The first step will be to envision the end state – that glorious nirvana where the cloud-enabled network is flexible, scalable, and responsive. Then the conversation can naturally turn to “how can we get there?” Top-level planners will play a strong role in mandating change, but as we’ve seen across our customer base, the hands-on-keyboard administrators need to be fully on board for any cultural change to take hold.
Prepare the battlefield
Meaningful cloud migrations – the ones that move beyond “lift and shift” to actual development and operation of cloud-based applications – cannot happen in a vacuum.
As a company which specializes in DNS – the back-end infrastructure which moves information through any network – we’ve seen plenty of organizations try to coordinate and optimize their cloud operations long after the systems are up and running. This is almost always a mistake.
That’s because the infrastructure piece of cloud operations can be bafflingly complex when implemented at scale. Tasks like managing all of the pathways between resources through conditional forwarding rules can consume enormous amounts of time and energy if the back-end isn’t set up properly to begin with. We certainly wouldn’t recommend trying to implement JEDI with out-of-the-box Microsoft DNS in the cloud.
In our experience, mapping out the architecture and infrastructure of a cloud deployment in advance pays significant dividends later on. Even better, if you can align the structure of your cloud around business objectives and operational goals, the entire enterprise will be set up for success in the long term.
Start planning now
As we’ve noted elsewhere, the JEDI cloud revolution probably isn’t going to happen all at once. While Microsoft is contractually obliged to get their solution up and running in a very short amount of time, actual adoption is going to be a trickle long before it becomes a flood. It will take time for top-level directives to filter down and get to the operational stage.
All of this means that agencies should really start planning now for how they’re going to implement the JEDI cloud before all of the mandates come raining down. While they have the luxury of time, DOD agencies should be thinking about their business goals and building the architectures they’re going to need moving forward.
That includes paying attention to basic infrastructure like DNS – the kind of back-end functions that administrators often take for granted, but will need to basically reinvent for cloud deployments at scale. Putting your agency’s house in order will make the process of migration to the cloud (not to mention the actual operation of those clouds) a whole lot easier.