Pentagon scraps $10bn JEDI winner-takes-all cloud contract The Pentagon has killed off the $10bn JEDI IT contract that Amazon, Microsoft, Oracle, and others spent years fighting over. The US government's Dept of Defense announced today the decade-long, single-vendor, winner-takes-all cloud deal would be scrapped. "With the shifting technology environment, it has become clear that the …
"MS will now spend the next 10 years fighting to get the contract cancellation revoked."
I suspect that the MS folks will probably be way too busy padding the bill they will send the government for work already performed and for contract termination costs to waste effort trying to get the contract reinstated.
Lowest common denominator is pretty amazing these days. Postgres, Docker, Kubernetes, nginx, HAProxy, rabbitmq, Kafka, Hadoop? These are the technologies dreams are made of, unless you're a cloud provider looking to lock in the government forever.
Yeah, go back and read the earlier articles on this smorgasbord of FAIL. The whole point was to deploy a solution with particular requirements reasonably quickly, and at a reasonable price. Microsoft won because they met those requirements. Amazon and Oracle lost because they literally fought to have the requirements changed instead of delivering what was needed, among other very good reasons.
This is nothing other than the egos of overfed swine trying to squeeze out a competitor from a feeding trough of government money. In a saner world Larry Ellison would be in jail and Oracle would be banned from government business for the systematic over billing and extortionate pricing of contracts the US taxpayer is stuck footing the bill for. Worst is that this was intended to address actual needs, and this fiasco has jammed up the works for what was to be a fast track project for a decade.
I don't see why you think one competitor is better than another here. It's likely the Trump-Bezos feud is what got AWS replaced with Azure.
But anyway. Your post reminds me that vendors don't just want technical lock-in. Emotional lock-in means you'll keep choosing them no matter what.
Sherlock Holmes icon in lieu of Doctor Phil.
"Postgres, Docker, Kubernetes, nginx, HAProxy, rabbitmq, Kafka, Hadoop? These are the technologies dreams are made of..."
Uhh, how many of those technologies have you designed, deployed, and supported through production at any kind of nontrivial scale? I believe I can answer yes on 5 of them and I'd characterise the experience as being more of a nightmare. nginx was the best, but that may have been because we weren't asking much of it; I think I only had to fix a few bugs to make it usable. rabbitmq and postgres, the less said the better. K8s, no. Just no.
Maybe this was meant tongue in cheek?
"Postgres, Docker, Kubernetes, nginx, HAProxy, rabbitmq, Kafka, Hadoop? "
Almost all of which require you, the customer, to have an infra team managing the legacy IaaS platform those things sit on top of. Replace Hadoop with Databricks and you gain a lot of efficiency for spark workloads because it's running as a service. Admittedly Azure has a managed Postgres, and maybe the others too, but could you guarantee that the versions are feature compatible? No, of course you can't, and neither can the cloud providers. That leave you installing it manually on a Linux VM at 8pm on a Friday while I'm out drinking because I decided to go with real cloud services and commit to vendor lock in in a good way.
Sure, it's harder to leave, but that's not as hard as maintaining two or three clouds so the juice is worth the squeeze. Anyone telling you different doesn't understand the question.
I don't quite understand your point. Saying "use postgres" doesn't mean "don't use hosted postgres". Only mentioning Azure, and calling non-MS "legacy" presumably makes you a MS-specific person, but there's no need to be threatened by Postgres.
The point is that if you're managing your own system images and infrastructure then you've lost all of the value of cloud computing. All of the clouds have managed offerings, and generally those are incompatible so if you're going multi-cloud then by definition you'll be using IaaS and legacy architectures which increase operational costs and complexity.
useful vendor-agnostic setup
A big problem with vendor-agnostic designs is that it makes it very easy to swap between vendors, and the vendors hate that. As a result they invariably implement the design (which tends to be lowest-denominator) as specified, and then add some "extensions" which make their solution more appealing to their customers, but which are of course incompatible with the additions from other vendors. The result is that the agnostic part is buried under a bunch of proprietary extensions which create the same vendor-lockin as before. Apple and Microsoft are past masters at the "embrace, extend" model.
Because jarheads make the very best systems architects. (Said with much love for those who did not serve in the Chair Force.)
Does the military build its own aircraft? Tanks? Guns? Shoes?
A JEDI-like system is one of the most complex things on the planet. It's doubtful that the military even has the chops to properly spec out such a system. (A fact that Amazon attempted to exploit to lock up the contract.)
Although its fashionable for governments to privatize and outsource there are core functions that cannot be left to a commercial provider. One important one is archiving, the preservation of information that's essentially the institutional memory of the country and its society. Since much of that storage is now in electronic form its important that governments control that storage and the infrastructure that serves it.
Our (US) government already has a significant investment in storage and computing through the NSA. This organization is dedicated to the security of the country but has wildly misunderstood the nature of threats to our society. It expends enormous resources capturing and categorizing every iota of communication -- that is, spying on everything and anything -- but what it should be doing is developing secure archiving for not just the government but for all of us. (It, along with the CIA, should also be taking a proactive role against malware instead of trying to develop it in the hope it could be weaponized against an adversary.)
We're learning the hard way that the price of convenience (which turns out to be not that cheap anyway) is a lack of resilience.
Who would deny that is an AWEsome opportunity for some folk ..... who both know what they are doing and how to do it with IT Command and Control and what needs to be done. Default on any of those requirements and failure is guaranteed ... swifter than a speeding bullet :-)
> You will never find a more wretched hive of scum and villainy. We must be cautious.
While I understand the sentiment wrt the goverment, in the context of this article, that quote would be better applied to the cloud providers:
PHB: "Let's do a request for tender for cloud services, send it to the usual big cloud providers"
BOFH: "You will never find a more wretched hive of scum and villainy. We must be cautious."
I suspect that Biden is going to be rifling through the couch cushions for every thin dime that can be freed up, but the sales of the F-35 aren't a big part of that. We will sell the damn things to almost anyone at this point, in the hopes of staving of the probably inevitable death spiral of the old flying turkey.
Every sale to a current(and perhaps soon to be former) ally offsets some of the planes the US government is forced to buy on their own dime. Planes that still don't meet the basic airworthiness and performance requirements of their specifications and contract. They never will. It was never designed to deliver on it's promises, it's just a bottomless pit designed to siphon off budget from the DOD.
Sadly, when bad actor vendors will happily cripple projects like JEDI even when their bids lose on their merits, it makes it hard to hold down costs or get anything done. So we will blunder into our next conflict under prepared and with less resources than we should.
"Sadly, when bad actor vendors will happily cripple projects like JEDI even when their bids lose on their merits, it makes it hard to hold down costs or get anything done.
Perhaps, if those bad actors spent less on litigation and more time as well as money, on ensuring the quality of their output, everyone would be happier.
As it is, I suspect they will milk any contracts they do win, until the cash cows are dried and wizened.
I would like to be a fly on the wall at Oracle right now!
Its about time the USA stopped throwing away money on so called Defence, when the true term is Offence.
When was the last time anyone invaded the USA? "December 7, 1941"
How many countries has the USA invaded or infiltrated in some way since ten?
1949 Syrian coup d'état
1949–1953 Albania
1951–56 Tibet
1953 Iranian coup d'état
1954 Guatemalan coup d'état
1956–57 Syria crisis
1960 Congo coup d'état
1961 Cuba, Bay of Pigs Invasion
1961 Dominican Republic
1963 South Vietnamese coup
1964 Bolivian coup d'état
1964 Brazilian coup d'état
1966 Ghana coup d’état
1967 US manufactured coup in Greece
1971 Bolivian coup d'état
1970–73 Chile
1980 Turkish coup d'état
1979–89 Afghanistan, Operation Cyclone
1980 -1988 material support for Iraq against Iran
1981–87 Nicaragua, Contras
1996 Iraq coup attempt
2001 Afghanistan
2011 Libyan civil war
2011–present Syria
IRA
Thanks for that one, now included, and you are probably correct but I cba to look. What I like are the downvotes because it proves quite a few people can't stand facts and the truth.
I get the same from Flat Earth cultists, AGW deniers and Creationists who are all basically the same in that they are biblical literalists and mainly 'Muricans i.e. not all Americans.
That scene had the potential to be perhaps the best ever. Unfortunately, the multi-season gap between the setup and delivery meant that the director felt compelled to add the flashback. I mean, I get it from a marketing/pr perspective, but it destroyed the art of that moment.
Comparing SQL Server from Microsoft to the serious and professional databases from Oracle and IBM, the idea that Microsoft and Amazon were competent to provide what the Pentagon needed but Oracle and IBM were not struck me as ludicrous.
But if it was clear that what the Pentagon was a buzzword-compliant cloud thingy, which Microsoft and Amazon tried to offer, while Oracle and IBM were instead trying to flog off their old-fashioned on-prem traditional database technology, they got what they deserved, right?
After that though struck me, though, another one came along. What if what Oracle and IBM were offering was something that would actually meet the operational requirements... and something using the new cloud style technology could not meet them or be made secure given the state of the fancy new cloud stuff now or in the foreseeable future?
Then it's the Pentagon that's getting what it deserves for turning up its nose at what would work from Oracle and IBM.
Then: Uncle Sam decides to give out a TEN YEAR contract for services in the fast-moving technology environment.
Now: "With the shifting technology environment, it has become clear that the JEDI Cloud contract, which has long been delayed, no longer meets the requirements to fill the DoD’s capability gaps,"
In the words of Homer Simpson: "D'oh!". It didn't take a genius to work out that even 'back then' technology changes fast, so the capacity for change should be built in... eejits, the lot of them.
> Microsoft focused "on our customer, and not politics or litigation,"
Amazing how virtuous you can be when you won the damn contract. Bet you many $ that had the contract gone the other way there'd still be the litigation, just with a different name attached. sed 's/Amazon and Oracle/Microsoft/g'
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
