milCloud 2.0: Dead Cloud Walking
The US Government, at least in the post-war period, was extremely effective at funding and executing some of the best technology research the world had ever seen. However, its track record at productizing those very same programs is much less impressive.
In an ideal world, the federal CIOs and CTOs would understand that when it comes to building and productizing technology, the private sector usually is much more efficient. However, the government has had to learn this lesson time and time again, often at the cost of dumping hundreds of millions of taxpayer dollars into projects it ends up later throwing out in favor of its commercial counterparts. The latest in a series of these examples is Defense Information Systems Agency’s (DISA’s) milCloud 2.0.
As late as 2017, DISA awarded an exclusive contract valued at almost $500 million to CSRA, LLC (later acquired by General Dynamics) to develop “milCloud 2.0 Phase 1 commercial infrastructure services in Department of Defense (DOD) facilities.” Essentially, DISA wanted to create its own platform to meet the demand for an elastically scalable cloud solution to provide secure and affordable general purpose cloud services to its Defense Mission Partners.
This requirement was deemed so urgent that by February 2018, DISA managed to push for an earlier release of milCloud 2.0 due to high demand. By the end of that same year, Forge.mil became the first DISA-managed application to successfully migrate to the new milCloud; by November 2019, DISA had integrated 30 of its unclassified agency applications to the upgraded platform.
Despite that many Defense agencies and DOD field operations (DAFAs) failed to meet their migration deadline for the end of Q4 FY 2020, a DISA official still reported that the workloads transferred to milCloud 2.0 had doubled since the 2018 order to upgrade.
But last December, DISA announced its plans to let the milCloud 2.0 contract expire this May, even though at one point it was running over 4,000 workloads distributed among some 90 different defense-agency clients.
So why is DISA now letting milCloud 2.0 die?
On the surface level, the agency simply decided that the requirements for the platform it no longer meets the DoD’s needs.
Looking a little deeper, it’s clear that DISA was eventually forced to face the music.
On LinkedIn, Enterprise Cloud Management Agency Director Paul Puckett wrote: “We have a REALLY big problem with continuing down a path that was the right decision when it started but the wrong decision now as the world shifts around us. milCloud made sense in 2013, but it doesn’t in 2021.”
The reality is that milCloud 2.0 was doomed to be a lackluster service from the start when compared with its private-sector counterparts. At the time of milCloud 2.0’s launch, private firms had already been in the cloud-technology game for nearly a decade; DISA’s physical hardware resources paled in comparison to the legions of servers worldwide belonging to AWS and Microsoft. Last year, GDIT even integrated Amazon Web Services capabilities into milCloud 2.0 so clients could “take advantage of the latest technology and innovate more quickly with artificial intelligence, machine learning, cyber sensing, and other emerging capabilities.” In short, GDIT, sensing the drift between the mission owner requirements and the MilCloud 2.0 contractual requirements and turned to AWS to provide MilCloud 2 customers with the latest cloud services because it did not make sense to do so its own. Amazon today manages 38 data centers worldwide, with each housing between 50,000 and 80,000 servers; DISA’s capacity, while vast in its own right, does not even begin to compare.
With respect to the future of DoD Cloud, DoD Leadership understands the times have changed. DISA’s “Strategic Plan FY2022–2024” explains that “[the] Joint Warfighting Cloud Capability (JWCC) seeks to create a multi-vendor acquisition vehicle that the greater DOD can leverage to obtain services directly from commercial cloud service providers. This would reduce the need for third party resellers, integrators, achieving efficiencies as a result.” The message reads clear: cut out the middleman, strengthen purchasing power of the DoD as a single purchaser, and leave the heavy infrastructure lifting to the private-sector.
Friends, Fellow Defense Contractors, countrymen, lend me your ears. I come to bury milCloud, not to praise it. The evil that systems do lives after them. The good is oft interred with their hard drives. So let it be with milCloud. The noble commercial clouds hath told you that milCloud was ambitious. If it were so, it was a grievous fault and grievously hath milCloud answer’d it. Here, under leave of the commercial clouds — for they are honorable clouds; come I to speak at milCloud’s funeral. It hath brought many workloads home to DISA whose ransom did the general coffers fill: did this in milCloud seem ambitious? When that the DECC applications have cried, milCloud hath wept. You all did see that on Black Friday I thrice presented the milCloud PMO a challenge coin which it did thrice refuse: was this ambition? I speak not to disprove what the commercial clouds spoke, but I am here to speak what I do know. You all did love it once, not without cause: what cause withholds you then, to mourn for it? O judgment! thou art fled to brutish beasts, and men have lost their datacenter ambitions. Bear with me; my heart is in the coffin there with milCloud, and I must pause til it come back to me.
