Pentagon Cloud Procurement, Deux

The U.S. Department of Defense last week announced a $9 billion contract to get the major tech companies to provide its next-generation cloud computing infrastructure and application platform. I think.

My hesitation comes from reading the DoD announcement, which appears to award a “hybrid (firm-fixed-price and time-and-materials, indefinite-delivery/indefinite-quantity) contract with a ceiling of $9,000,000,000” to Google Support Services, Oracle, Amazon Web Services, and Microsoft each. The awards have sequential contract numbers, so the $9 billion bandied in the press really looks like $36 billion.

This announcement is a far cry from the original JEDI contract circa 2018 (depending on how you count the time in litigation) that was worth two billion bucks over 10 years, which was won, and lost, and won by almost every company receiving the awards this time. The survivor of that bidding and litigation process was Microsoft which had a nanosecond to celebrate before the pentagon pulled the plug and went back to the drawing board.

At the time, I suggested that there would need to be another procurement process, that a single vendor selection would not fly, and that the new approach would force the major vendors to achieve a level of interoperability heretofore not seen.

My analysis was based on history, not reporting. Large government procurements are a good way to bring order to emerging industries by forcing them to develop standards if the individual contributors want to play. Look no further than SQL and the rapid evolution of the relational database for an example.

Process Under Wraps

So here we are, another process that was kept largely under wraps (hopefully out of court), and multiple vendors received serious awards to put up or shut up. At the time of the last award, I also suggested that a single vendor would emerge as the sole or prime contractor dividing the work among its peers. There is nothing that I’ve seen that anoints one of the vendors first among equals, and that is concerning.

My concern stems from a belief that this award will necessitate vendors to develop areas of relatively superior knowledge and expertise. Without someone calling the shots, there might be lots of duplication and waste. But we can hold that discussion for another time once we see how things begin rolling.

The whole relative superiority thing is serious because it may signal a fracturing of the industry along new lines with significant repercussions throughout the tech world.

Specialty Applications Analogy

To make an analogy, James Watt invented the modern piston-driven steam engine. It was good at all sorts of things, like driving pumps to keep coal mines dry enough to work in.

But you know how things go: somebody invents something, and others develop increasingly specialized applications. Eventually, steam engines were used successfully on ships — both riverboats and oceangoing freighters — and in railroads. Moreover, as time passed, fewer businesses elected to do both kinds of transportation because they had their hands full just doing one.

Piston-driven steam engines died out a long time ago, but we still depend on steam turbines for things like electricity generation.

That’s just one example. In software, there are many specialty areas where vendors may have competing offerings, such as databases, security, applications, development platforms, and specialty hardware for accelerating database operations. My bet is that the specialization I suggest will happen along those lines.

My Two Bits

In all of this, it seems to me that Oracle and Microsoft might be the companies with the biggest decisions to make because their businesses are strewn all over the place, from databases, languages, and middleware to cloud infrastructure and much more.

Does this mean everyone else’s offerings will die on the vine? Probably not, but it may logically force everyone to innovate a new generation of products for the commercial markets. Also, some of the innovations for this project will likely leak into the public sphere anyhow.

The idea of buying from a single vendor, much as each vendor might like it, is rapidly receding to the background, and today application support looks a lot like the insides of a PC with hardware from a large community of vendors providing final assemblers with an array of choices.

Because this is the Department of Defense, the consortium will need to be extra, extra, extra careful about the seams and handoffs because those can be the weakest areas most attractive to hackers. Everybody knows this.

For now, regardless of how this plays out for the DoD, expect a significant transformation in how we do software in the years ahead because tech is losing its revolutionary status and becoming just another part of the economy.