Last week the Pentagon might have awarded its $10 billion contract for cloud computing to Microsoft. The program goes by the acronym JEDI for Joint Enterprise Defense Infrastructure and it has been attracting vendors like a dog attracts fleas for several years. It has been marked by fierce litigation too, so the award may not be the end of it.
Oracle sued over the process claiming it was unfair and focused its ire on Amazon which appeared to be the leading candidate. CTO Larry Ellison made a point of showcasing the benchmarks of Oracle’s Autonomous Database against Amazon at multiple OpenWorld conferences to show Oracle’s technical dominance. Amazon had been making a market share argument of the variety that we have the most datacenters so we’re the best. But cloud computing is a 3-leged stool and you need infrastructure, software and tools to really be a contender.
Then Donald Trump got involved as only he can as the Commander in Chief. There’s no love lost by Trump with respect to Bezos who also owns the Washington Post which Trump dismisses as the Amazon Washington Post in not so sharp distinction from his description of the failing New York Times.
According to an article on CNBC from Saturday, “An upcoming book on James Mattis’ tenure as secretary of Defense claims President Donald Trump told Mattis to “screw Amazon” out of a $10 billion cloud contract for the Pentagon.” Mission accomplished, I guess. If true, that looks like self-dealing rather than putting the concerns of national security first and foremost, but I digress.
But more than all of the gossip about who gets the contract there ought to be questions about what the contract will do to shape the cloud industry. Where to begin?
Microsoft and who else?
First, is Microsoft the only contractor in this deal or is it the prime? The difference could be big. If the company founded by Bill Gates and Paul Allen is the prime it leaves the door open for the losers in the derby to make strategic contributions. For instance, it defies logic to think that the Oracle Autonomous Database would be excluded from delivering its functionality for national defense. In my humble opinion it is the most performant and secure thing in the cloud, and I say this not to dispute other products’ greatness but only to point out that Oracle has delivered a hardware and software solution to market that achieves that purpose.
But this brings up a thorny issue. To get the most out of Oracle’s technology you really have to buy its Exadata storage system and maybe some other pieces. It’s a standard part of the Oracle cloud system now being deployed but it might raise concerns about vendor lock-in for a customer that likes to have no single point of potential failure. For Oracle to get a piece of the contract it might have to bend a little by making its solution more open.
That’s not far-fetched. Over the summer Oracle and Microsoft have been cozying up with grand strategies to make their systems more open to each other. A coincidence? I think not. It’s interesting to me that Oracle quit making headlines about its JEDI lawsuit after August. More interesting is that as much as Trump wanted to “screw” Amazon, he didn’t seem to offer help to his friend Safra Catz who is now CEO of Oracle.
The cloud computing utility
I’ve written frequently that the IT industry is commoditizing and that a utility is forming. Like the electricity industry in which many vendors adhere to accepted standards and present the appearance of a continental grid (it’s not a single grid though but that’s a story for another time), an IT or cloud computing utility might look similar.
If all that’s true then Microsoft sits in position to become the standard setter and gate keeper. There would be a lot of contention over databases but perhaps the Oracle product might be used in large, ultra-secure situations. Maybe Oracle and Microsoft would also agree on the central importance of Exadata as the hardware foundation too. There are lots of permutations but most importantly JEDI might force a truce and foster better interactivity standards, some of which are clearly underway.
In that kind of environment there might also be room for IBM’s Watson and Amazon’s datacenters as well as other advanced technology.
My two bits
I can’t fathom how the award is the end of the story. In a lot of procurements, the government puts itself in the position of choosing the best and most viable vendor and of telling all the others the equivalent of your baby is ugly. That’s not the case here. America brims with technology talent and products (a consequence of the government sponsored space race, ultimately) so each contender could have done a credible job of at least being the prime contractor.
The big impact of the deal will be on how it shapes cooperation among the oligarchy that now controls much of our technology future. In the 1980s the CIA bought relational database technology from a startup named Oracle and a relatively young Larry Ellison became its chief support person, at least for a time. Relational databases were much better at managing data and reporting than the flat file systems then in use. Even Oracle’s company name suggests being able to predict the future in the way that the ancient Greek Oracle at Delphi did.
That win spawned the RDBMS revolution, we all became SQL fluent and Oracle never looked back. The JEDI contract feels like one of those moments when a government procurement can change an industry. That’s why it was so closely contested and why I think we’ll now see the industry coming together and creating new interoperability standards.
So now Oracle is appealing the Pentagon $10 billion JEDI contract to provide cloud computing solutions, which was awarded to Amazon. A release on PR Newswire states,
“The Court of Federal Claims opinion in the JEDI bid protest describes the JEDI procurement as unlawful, notwithstanding dismissal of the protest solely on the legal technicality of Oracle’s purported lack of standing. Federal procurement laws specifically bar single award procurements such as JEDI absent satisfying specific, mandatory requirements, and the Court in its opinion clearly found DoD did not satisfy these requirements.”
At the same time the DoD announced an official review of the award saying among other things
“We are reviewing the DoD’s handing of the JEDI cloud acquisition, including the development of requirements and the request for proposal process. In addition, we are investigating whether current or former DoD officials committed misconduct relating to the JEDI acquisition, such as whether any had any conflicts of interest related to their involvement in the acquisition process.”
Frankly, a lot of people probably (likely, definitely) don’t care about JEDI and what it means, but if we were all small-d Democrats in this republic, we might (ought to) care. Here are some reasons.
Advancing the technology
Throughout the post war history of tech and its relationship with government there’s been a great symbiosis that has advanced the frontiers of technology while giving government better, faster, and cheaper solutions. Government’s investments in numerous companies and think-tanks drove research and development R&D and supported emergence of globe-leading new companies and millions of employees who pay taxes. All in all, it has been a productive relationship and a good investment.
But now the Pentagon wants to chuck all that and go with a single provider that it hopes will continue to advance cloud computing and, ironically, it can be argued vociferously as Oracle continues to do, that the JEDI process has not selected the strongest technology for the job.
For instance, Oracle founder and CTO, Larry Ellison, takes no small measure of glee when he presents benchmarking data at Oracle OpenWorld (coming to San Francisco in September) comparing Oracle’s technology with Amazon’s. By many measures, it’s just not that good for huge deployments though many Oracle users use Amazon products for sandboxes in their development labs because it’s less expensive. But if you’re offering $10 billion take it or leave it for what amounts to the kitchen sink, might cost suddenly move to the back burner?
But the Pentagon’s concern about having multiple vendors in the soup for what will be its cloud computing foundation at least to the middle of this century, is also understandable. Cloud computing, heck computing in general, is still a Balkanized jumble of vendors with often uncooperative standards that prevent information sharing. It’s getting better but it is still unacceptable when you’re dealing with technology that will be used to make split-second decisions about friend and foe, war and peace.
I get that.
But it’s no reason for a sole source award. With history as a guide we should be encouraging a multivendor award that demands that vendors do a whole lot more to advance inter-connectability. Moreover, that’s the direction this industry is moving in anyhow, i.e. toward a big information utility much like the electric utility we all use. The JEDI award should be used to accelerate this process.
Don’t be confused, there is no single vertically integrated electric utility from sea to shining sea. Instead there are thousands of companies and solutions that work within the same standards so that there’s an appearance of a utility and a grid. That’s what the JEDI procurement should be working towards.
Sole source award and Oracle doesn’t have standing?
Okay, I’m not a lawyer but you have to admit that a ruling that Oracle doesn’t have standing in the appeal seems at best contrived. This procurement is a gnarly, complex can of snakes that about 1,000 people on the planet truly understand and virtually all of them work for DoD, IBM, Microsoft, Amazon, Oracle and a few other places (China, Russia), but essentially the companies that competed for the business in the first place.
So if Oracle doesn’t have standing, who does? Anybody? And is that any way to run a democracy? The PRNewswire piece also says that
“The opinion also acknowledges that the procurement suffers from many significant conflicts of interest. These conflicts violate the law and undermine the public trust. As a threshold matter, we believe that the determination of no standing is wrong as a matter of law, and the very analysis in the opinion compels a determination that the procurement was unlawful on several grounds.”
Bragging rights and picking winners
Aside from the $10 billion (plus overages) that some lucky vendor will gain from this award, there’s the matter of bragging rights in the marketplace. It’s hard to believe the successful vendor won’t try to put some marketing muscle behind claims of superiority thanks the endorsement such an award provides as a natural consequence. That means picking winners, something that politicians raged against when the Obama administration made an investment in a solar panel maker.
That’s why the after award litigation is so hot. Oracle has the resources and is known for being a persistent litigant when it feels wronged and, in this case, it’s hard to see how they could not challenge the award.
This litigation could tie up and real work on the contract for years and deprive the Pentagon and the American people of advances in defense and security that they should expect from their government. It’s not too late for the Pentagon and DoD to take a step back and re-evaluate their goals.
My two bits
JEDI should be thrown out. It’s not a stretch to say that the better technology was not selected, at least by some important measures. Also the belief in a sole source was misguided at best. But even if you throw JEDI on the ash heap of history, you don’t have to begin again from zero. A broader requirements set based on a demand for interoperability could be put forth and several other vendors (all of them American companies) should be incorporated into the process. It’s hard to believe that Microsoft, with its expertise in system software, networking, apps, tools and databases wasn’t included. You can say similar things about IBM and HP both of whom have significant hardware, networking and Unix chops. And don’t forget IBM’s Watson. Add Oracle to the mix for their autonomous database, fault tolerant and ultra-secure hardware, apps and more.
The trend so far, to me, suggests that there are too many Ids and egos involved in this deal and not enough logic and dispassionate decision-making. The American people deserve better as does the planet whose fragile peace depends on it.