Oracle goes on for 50 pages about why it thinks the Pentagon's $10bn JEDI cloud contract stinks Ahead of its first day in a US federal claims court in Washington DC, Oracle has outlined its position against the Pentagon's criteria for the award of the Joint Enterprise Defense Infrastructure (JEDI) cloud contract - which, crucially, Oracle does not currently meet. Big Red's lengthy filing questions the basis of Uncle Sam' …
IMHO, the whole Federal Government, nay, politicians everywhere, could use more of a sense of humor. If bureaucrats were able to laugh at themselves more, maybe politics would be less vicious. Hell, if people were able to laugh at themselves more, the world would be a better place.
Also, obStarWarsReference: Begun the cloud wars have.
IMHO, the whole Federal Government, nay, politicians everywhere, could use more of a sense of humor.
There was a program a year or two ago which pondered why England/Britain had never really had a bloody revolution like some of our European neighbours. The theory put forward was that as satire and taking the p**s out of those famous or in power was in our blood, the top brass never managed to completely disappear up their own arses here.
Thats silly because Britain DID have a bloody revolution. One of the first monarchical overthrows in the western world, in fact. The English Civil War was INCREDIBLY bloody and deadly and part of the reason why the thought of a standing army was repugnant for so long and why and public use and carry of weapons is socially anathema. Over 200 000 people died in less than a decade. A higher percentage of the population died in the civil wars than in World War 2.
That's why we "didn't have one" - because we did in fact have one, earlier than others, and decided that that shit wasn't for us.
>That's why we "didn't have one" - because we did in fact have one, earlier than others, and decided that that shit wasn't for us.
And of course with Americans being so shitty at history we had to one up it much later with 600,000+ dead in just a few years. Sadly with so many being so ignorant of history might even have another someday. Sure as shit didn't get rid of the arms or the standing army.
Except that 'The English Civil War', otherwise the War of the Three Kingdoms, wasn't actually a revolution a la France et al.
A lot of it was actually about the burgeoning middle class acting like the kids and rebelling against the Aged Parents; part of it was about religion, mainly the Irish Catholic Confederacy revolt, but you could also include the introduction of the new Prayer Book in Scotland; and some of it in the earlier phase of the English part was specifically *not* about replacing the monarch, just his advisors.
because we did in fact have one, earlier than others
And the net result was a semi-theocratic semi-monarchy-in-all-but-name that pretty much everybody except those who profited from it hated[1]. And which collapsed as soon as Cromwell died because his son didn't have Ollies brains, charisma or ruthlessness.
[1] The hardline protestants hated it because they didn't kill enough catholics, the catholics hated it because they were being killed, the ordinary people hated it because fun stuff (dancing, strong drink etc etc) got banned and the aristocracy hated it because the people running the country were not the Right Sort.
"IMHO, the whole Federal Government, nay, politicians everywhere, could use more of a sense of humor. If bureaucrats were able to laugh at themselves more, maybe politics would be less vicious"
Thats all well and good, but national defense isn't funny and shouldn't be treated as some light hearted activitry. There's a time and place for levity - this isn't it.
JAWA
What - a dodgy Czech-made 2-stroke motorbike from the 1980's that'll fall apart as soon as you try to use it? The one where the concept of quality control was even worse than that at Lucas Electrics?
(OldestBrother had one. It was truely, truely crap. But it was cheap, very, very, very cheap)
Oracle's filing said that US "warfighters and taxpayers have a vested interest in obtaining the best services through lawful, competitive means... Instead, DoD (with AWS's help) has delivered a conflict-ridden mess in which hundreds of contractors expressed an interest in JEDI, over 60 responded to requests for information, yet only the two largest global cloud providers can clear the qualification gates."
Sounds like sour grapes if I'm not mistaken.
The fight for JEDI will be hard and horrible, and lots of shenanigans can be expected.
if that supplier were Oracle.
The number of points of failure is completely irrelevant, what is relevant is that oracle has lost that is what really matters to them, never mind that the competition is better than they are.
oracle is bad loser!!!!
I don't think Oracle has ever claimed they should be the sole winner. They just don't like that a sole winner was chosen, since it will be such a big advantage for the cloud provider who gets it.
I tend to agree, the DoD "stamp of approval" on Amazon as the sole source for this contract will legitimize them with other government agencies and they may grow so big it may be impossible for other cloud providers to compete for non-government contracts due to their advantages of scale.
There's no reason they couldn't have split up the contract several ways. Though I think it would be funny if they did and Oracle still wasn't one of the winners!
> They just don't like that a sole winner was chosen, since
...since they knew it wouldn't be them, shutting them out of the money entirely. If they thought they might end up with the whole pie, it would have been perfectly fine with them to have a single winner-take-all contract.
They're going to get a shit load of grief from Oracle for not giving them part of the contract. Can you imagine the two or three orders of magnitude more the DoD would get from Oracle if they had let them have some of it?
DoD surely knows what a clusterfsck Oracle produces on most big jobs...
The single vendor issue is interesting.
One thing it maybe makes easier is interoperability. Even within a big organization, one gets champions of this or that format, or way of doing things, and when you want to combine systems, this effect can make it far more difficult or cost ineffective than if one overarching vision had been implemented.
(See The Cathedral vs the Bazaar perhaps).
Now I can see where perfect interoperability can be a bad thing too - think of all the data collected on us by various of the big tech outfits, 5 eyes, law enforcement and so on being perfectly coordinated.
Definitely useful if you can do it to your adversary, but not so nice when turned on your own folk, who are increasingly being treated as the adversaries by governments.
Maybe not to worry - the effort in writing a good standard - and getting it signed off on is so high it's mainly given lip service and skipped as a high cost line item that won't pay out anyway.
As a contracting consultant, I was told that insisting on agreement with design details pretty thoroughly before starting work caused some pain in the customer - but then lauded as our results were always top notch. It took awhile for someone to connect the dots that having the good plan was the hard part, and just hardware and software design was then a fairly trivial exercise.
Having also worked with DoD, I don't think that this particular bulb will light up for them.
And then there's the XKCD on standards we've all seen....
Which separately reminds me of the numbered jokes in the old folk's home - just say the number and everyone laughs (speaking of not "getting" the ideas of timing and presentation).
927
While a $1 billion/yr contract is nothing to scoff at, I think it's a bit naive to assume there aren't similar sized contracts, if not larger contracts, from the 3-letter agencies. This is *ONLY* for the DoD. There is plenty of pie to go around, and I would imagine that's been taken into account.
Case in point (you can find countless with a little bit of searching):
https://www.seattletimes.com/business/microsoft-says-it-wins-lucrative-cloud-deal-with-intelligence-community/
Granted, Amazon has absolutely been the leader which gives them a bit of a leg up, but this is just sour milk by Oracle. It might be a different story if MS were doing the same, but Oracle is just desperately looking for anything they can to prove that their cloud strategy isn't just a me-too joke (which it is).
My understanding of the key differences between the AWS and Azure proposals versus the IBM and Oracle proposals was that IBM and Oracle expected the DoD to pay for new facilities and wait for the facilities to be built while AWS/Azure had GovCloud facilities in suitable locations already and was expanding independently of the JEDI contract resulting in a significantly faster migration.
I'm not disputing that his gave AWS/Azure a significant advantage, it also gives the DoD a significant cost saving (DoD "cloud" spending is around US$1.2bn in 2018 and predicted to reach US$1.5bn in 2020 if JEDI is not implemented - JEDI is expected to bring cloud spending under US$1bn/year and US$10bn over 10 years which is around half the forecast amount of not implementing JEDI).
Is JEDI good value for the US taxpayer? Yes. Is it the best value? Maybe not.
Could Oracle provide a cheaper or better solution? Based on the evidence the DoD has previously hinted at where legacy DC providers have charged premium rates AND nickle and dimed the DoD for every minor issue they encounter it is unlikely.
JEDI is a result of the incumbent providers not being willing to compromise and control their costs over multiple reviews going back almost 20 years.
"Single point of failure?"
This is Oracles view on the contract, but not necessarily the reality.
There are likely to be 30+ data centre suppliers that the DoD can use, possibly even including Oracle if it avoids burning all of its bridges with this legal action. My personal view is that the legal action is a rear guard action to avoid Oracle revealing a huge revenue hole before it exits the cloud/hosting market.
If you need to set up a high availability system utilising GovCloud, the connectivity is in-place to get you into multiple locations within AWS or across DoD GovCloud providers. While the preference is likely to be AWS/Azure for the majority of systems requiring this level of availability, there is nothing to stop alternative vendors being used if they meet the DoD's criteria.
The real issue is that Oracle will have to be content with it's existing share of the DoD pork barrel, knowing that the barrel is shrinking rapidly so that they now have to be competitive with Azure/AWS in future tenders. Given that Oracle benefits from significant licence purchases from the DoD, MS and AWS are likely to see potential ways of saving the DoD money while preserving their own margins...
I think this has more to do with Oracle tech being removed by the federal government, i.e. migration from Oracle DB to Redshift etc.. Oracle sees a large portion of their revenue from federal contracts, and if this paves the way to migrate away from the Oracle tech stack for the military it may bleed over to other government agencies as well, I would say Oracle is probably fighting for its life here as the future doesn't look so great at the moment!
First and foremost, this is purely about ensuring compliance with DoD spending guidelines - DoD cloud spending has been significant and the DoD want to ensure this is brought under control:
https://assets.bwbx.io/images/users/iIyChnKQ1lHc/iODjz9dkexBE/v0/620x-1.png
Redshift is likely to be a future direction - historically Oracle gave friendly licensing terms to the US government that resulted in a lot of Oracle applications and databases being deployed because all users were effectively already licensed. Over time, those terms have become a lot less generous but cleaning up the environment to ensure appropriate systems are used rather than Oracle everywhere HOWEVER the DoD do have legitimate requirements for Oracle applications and databases so it won't completely disappear.
This post has been deleted by its author
I did a very small amount of research on this. The government justification for sole source seems rather hazy. But I'm guessing that they prefer the risks of being locked into a single contractor until the end of time to the risks of ending up with a digital Tower of Babel where every organization with a procurement office has a different contractor and a different -- probably incompatible -- cloud interface.
Anybody actually know anything about this?
It's obvious why this specific contract is sole-source.
It's to provide a leading-edge cloud service, not develop technology or standards, or figure out what is the lowest-common-denominator service.
If they had multiple suppliers, there would have to be compatibility. That means any one of the winners could halt cloud deployments until they had something compatible and working. A technology-trailing contract winner could effective make the service pointless. Sure, they wouldn't make money, but if your goal is preserving your very profitable legacy services through other contracts, undermining this contract would be an easy call.
That is precisely why there is only one winner. The other option for the DoD is to write an open-source API rule set that everyone can use if there isn't a suitable one already available.
Then Oracle would just refuse to use it which wouldnt be a bad thing in the long run.
From my research, the DoD is looking to consolidate it's existing ~300 data centre environment to something in the order of a few key providers (likely AWS and Azure given Google employees reluctance to work with the DoD and Google still having certifications to complete to allow them to provide a full range of services), approximately 50 state data centres to meet local state requirements and a few specialised data centres where required (i.e. Lockheed Martin have a few for F35 support).
The attempt to rationalise DoD data centres has been on-going for some time, but big increases in "cloud" spending (approx +$200M/year over the last 3-4 years) have meant that rather than trying to retrofit existing data centre arrangements, moving towards an AWS/Azure-type solution allows the DoD to control some of their costs while ensuring a higher level of compliance. Current cloud spending is around $1.2bn in 2018 and JEDI is expected to cost $1bn/year for 10 years.
In terms of lock in, it's an existing issue with many of their current DC's. While the theory is they can walk in and move all the hardware to another location, it likely just means they pay for everything twice as many of the existing contracts are not flexible.
Then EVERYONE involved when hiring this joker should be fired immediately. "Why do you want to leave your current employer" is the FIRST question I get in 80% of interview if I am currently employed. And by that, I mean that if I am interviewed by five people at one company, four of them will ask.
Federal procurement, especially around the DoD & WD, as always been a cesspit, but things have gotten out of hand.
Not that I would want to trust Oracle for a single $80 hammer.
As others have said, this is basically Oracle whining about not being awarded part of the contract (they know there's no way they could get the whole thing).
At one time Oracle was king but those days are long gone. These days, Oracle is only marginally better than Sybase (which basically hasn't seen much development since SAP took them over).
Oracle has partnered up with Microsoft in the DOD JEDI contract and they are also suing to win the award not by merit but by legal maneuvering and wrangling.
Oracle has no track record running a cloud server as big as Amazon AWS or Microsoft Azure. They are not simply qualified as evidenced by their loss in the first round of bidding. Oracle has become irrelevant.
Oracle's hope is to get Amazon disqualified so the big will go to Microsoft.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
