Oracle Challenges Pentagon's $10 Billion Cloud Computing Contract (theregister.co.uk) 101
Oracle has filed an official complaint with the U.S. government over plans to award the Pentagon's lucrative cloud contract to a single vendor. Rebecca Hill writes via The Register: The Joint Enterprise Defense Infrastructure (JEDI) contract, which has a massive scope, covering different levels of secrecy and classification across all branches of the military, will run for a maximum of 10 years and is worth a potential $10 billion. In spite of this pressure from vendors and the tech lobby -- as well as concerns from Congress -- the US Department of Defense (DoD) refused to budge, and launched a request for proposals (RFP) at the end of last month. Oracle is less than impressed with the Pentagon's failure to back down, and this week filed a bid protest to congressional watchdog the Government Accountability Office asking for the RFP to be amended.
In the protest, the database goliath sets out its arguments against a single vendor award -- broadly that it could damage innovation, competition, and security. Reading between the lines, it doesn't want either of Amazon or Microsoft or Google to get the whole pie to itself, and thus endanger Oracle's cosiness with Uncle Sam. Summing up its position in a statement to The Register, Oracle said that JEDI "virtually assures DoD will be locked into legacy cloud for a decade or more" at a time when cloud technology is changing at an unprecedented pace.
In the protest, the database goliath sets out its arguments against a single vendor award -- broadly that it could damage innovation, competition, and security. Reading between the lines, it doesn't want either of Amazon or Microsoft or Google to get the whole pie to itself, and thus endanger Oracle's cosiness with Uncle Sam. Summing up its position in a statement to The Register, Oracle said that JEDI "virtually assures DoD will be locked into legacy cloud for a decade or more" at a time when cloud technology is changing at an unprecedented pace.
Re: Trump files complaint against treason noose (Score:1)
Donald TRUMP cut in line at the supermarket!
Oracle might actually have a point here. (Score:5, Insightful)
I fundamentally dislike Oracle. Its an exploitative company that functions purely on ensnaring companies into deals that are far too costly then using legal shenanigans to stop them to leave.
BUT, they are right here. Giving the whole contract , all ten billion of it, to a single contractor (And lets be clear here, its either AWS or Azure. Google are capable, but they dont have the govt mojo to compete in this space) is straight up monopoly building, and it creates a single point of vunerability to the DODs systems. By splitting things up over multiple providers, it enhances competition, and divides up responsibility in a way better suited to national security.
And after all, they could still write "NO ORACLES ALLOWED" in it, right. (Well probably not, but hey)
Re:Oracle might actually have a point here. (Score:5, Insightful)
Oracle's only problem is that they are not that single contractor.
Re: Oracle might actually have a point here. (Score:3)
I have never seen Oracle complain when they were the single party awarded a contract, and there are plenty of times where this has happened. Perhaps the best that could be said of them is that they realize how many times they have screwed over the other side in such contracts and that this is a bad idea, but I suspect the real reason is that they are pissed that they do not get to be the ones to screw over the government this time. They just hate the player, not the game.
Re: Oracle might actually have a point here. (Score:4)
This is true, but that makes this all the more significant of a proof point of the value of competition. In a competitive landscape, there's going to be *someone* to call someone else on their shenanigans, even if it another usually bad actor.
Re: (Score:2)
Two sides to that (Score:5, Interesting)
I suspect the benefit to splitting things up may be obvious enough that I don't need to state it. On the other hand, over the years I've put a lot of thought into why companies use these clouds, and particularly AWS.
Years ago I developed a small private cloud using a lot of technology I designed and architected myself, with coding help from my employees and a contractor for the UI. It was mostly about storage, and some really nifty ways of managing virtual machines, but the main cost was storage. Multiple people asked me why we didn't use AWS for storage, so even after I had already looked into AWS I double checked a couple more times. What I found was that their storage was MUCH more expensive than some very solid, very flexible storage built from standard open source Linux storage components (cLVM, etc) and some 16-bay Supermicro chassis. AWS was super expensive for storage, and for virtual machines. So why are so many companies using them so much? Years later, I think I have a couple of answers.
There are a few reasons, but one is the level of integration of advanced things like auto-scale groups. Even getting just a load balancer working PROPERLY and configuring a static cluster of web servers is tricky normally. More often than not, the server clusters I see people deploy aren't actually clusters at all. They are a screwed up hybrid of a true cluster and a bunch of independent mirrors, which breaks things. AWS gives you a solid cluster in a few clicks. You can the easily save your entire cluster setup to your git repo as a Cloud Formation template.
The big clouds aren't the best way to get storage, they aren't the best way to run virtual machines, they aren't the best way to run databases. The magic is the integration - with a few clicks you have all the right DNS entries pointed to your new cluster of web servers, which talk to your DB cluster through the Lamda functions, all backed by the magic storage in a seamless way. With a beautiful API for programming it all. That's where the value is, how Amazon brings all these different things together seamlessly.
Breaking your operations up across a bunch of cloud providers meana giving up this seamless integration, duplicating whole data centers to another physical location with a few clicks, and haing everything still work.
If you're not going to take advantage of how everything is put together, you may as well save a few bucks and have a rack full of Supermicro gear on premises.
Re: (Score:2)
It depends on economy of scale. For a small company you are probably correct. As a larger, global company with multiple datacenters around the world we have done the cost comparison multiple times (about once a year for 5 years) and on-prem/private cloud was cheaper every time.
Re: (Score:2)
It depends on economy of scale. For a small company you are probably correct. As a larger, global company with multiple datacenters around the world we have done the cost comparison multiple times (about once a year for 5 years) and on-prem/private cloud was cheaper every time.
That's been my experience as well. I know for a fact and experience that large, global companies like NTT, ECB or CBS have hundreds of thousands of systems (computers, routers, gateways, etc) in-premise, their nature and business rationale being such taking them to the cloud makes no sense (financially and operationally.)
For some ops, it totally makes sense to go to the cloud.
For others, it totally makes sense *NOT TO*.
Three person company, scales from there (Score:2)
I started my cost comparison based on a three-person company renting a quarter rack (11U) duplicatds in two data centers. 6U was 32 hard drives, 3U was CPU, and the remaining 2U was the network switch and IP KVM.
Even at that level AWS was much more expensive, mostly due to the man power of occasionally maintaining it. Economies of scale make your own hardware cheaper as you scale up.
Re: (Score:2)
>> I started my cost comparison based on a three-person company renting a quarter rack (11U) duplicated in two data centers. ... 32 hard drives
> you had a non redundancy
Let me guess, with your keen eye for detail, you're a Bernie Sanders supporter?
What exactly is confusing to you about "DUPLICATED in TWO data centers"? How do you think 32 drives per unit would be configured? Perhaps as a RAID, aka Redundant Array of Independent Drives?
Re: (Score:2)
The reality is public cloud is easy, relatively reliable, and despite your assertion, it is usually significantly cheaper than on-premises, I noticed you didn't include rack space, building power, air con, maintenance, staff etc to run the hardware when you mentioned your comparison, when you do unless you are incredibly efficient it is nearly impossible to match AWS or Azure prices.
On top of that you've got to deal with infrastructure lifecycle, dealing with EOSL appliances, refreshing your infrastructure and phasing out the old ones, upgrading hypervisors, and proposing budget for those things to managements who don't understand that infrastructure lifecycle is a never ending process, who don't understand unless you sync their business projects the year before so that we can invest in hardware before hand, hardware delivery would take 4 to 6 weeks With cloud, assuming it's applicab
Re: (Score:2)
Re: (Score:2)
Why do you not question the idiocy of entrusting classified assets to a public cloud?
that's a different thread. Trusting 2 worse than 1 (Score:2)
That may be a different thread since this article is about whether to put it on one cloud or two.
I suppose putting classified info on two public clouds is twice as dumb as putting it on one public cloud.
Re: (Score:2)
And putting classified assets into any cloud is just asking for it. Don't let this one-or-two-vendor legerdemain distract you from the central question.
Oracle is weveel. (Score:3)
Re: (Score:2)
Agreed, although Oracle wouldn't be complaining if they were the ones getting the contract and they actively encourage customers to get locked in to their products so it's extremely hypocritical of them.
There are several frameworks allowing use of multiple cloud providers and easy migration between them.
Re: (Score:1)
In case you aren't aware, there is no such thing as a 10 year contract when it comes to the government. Because spending has to be re-authorized every year by Congress, all contracts are 1 year in length with an option to renew each year. Any corporation that does business with the government is aware of this. So each year it is theoretically possible for a different lobbyist firm or a different Congress to switch the contract. So even if the DoD does pick a single vendor for the engagement initially th
Re: Oracle might actually have a point here. (Score:2)
If you go single supplier you can get use their respective PaaS offerings easily. If you go multiprovider you end you either splitting your operations into one pile or the other to get paas or you end up going least common denom and end up shitty Iaas everywhere.
The choice, as others have said, will be aws or azure. Both allow an on-prem stack as well so you can control the physical as well as virtual.
Id really rather never touch an oracle db ever again if i can.
Re: (Score:3)
I fundamentally dislike Oracle. Its an exploitative company that functions purely on ensnaring companies into deals that are far too costly then using legal shenanigans to stop them to leave.
BUT, they are right here. Giving the whole contract , all ten billion of it, to a single contractor (And lets be clear here, its either AWS or Azure. Google are capable, but they dont have the govt mojo to compete in this space) is straight up monopoly building, and it creates a single point of vunerability to the DODs systems. By splitting things up over multiple providers, it enhances competition, and divides up responsibility in a way better suited to national security.
And after all, they could still write "NO ORACLES ALLOWED" in it, right. (Well probably not, but hey)
I'm on the fence here. I do think Oracle is in the right (Obi Wan's "from a certain point of view.") Such a 10B monopoly cannot be allowed to happen.
OTH, the benefit of having one cloud provider is seamless integration and scaling. If there are multiple contractors, then that will entail multiple providers, multiple cloud technologies, etc.
So the entire benefit of going to the cloud goes *poof*. If you (the generic "you") go to the cloud, you want to pick one provider, know the prons and cons and make
Re: (Score:2)
I'm on the fence here. I do think Oracle is in the right (Obi Wan's "from a certain point of view.") Such a 10B monopoly cannot be allowed to happen.
AWS [cnbc.com] and Azure [techcrunch.com] each do over $20B a year (and growing), adding $1B a year more to either one will not create a monopoly.
Re: (Score:2)
I'm on the fence here. I do think Oracle is in the right (Obi Wan's "from a certain point of view.") Such a 10B monopoly cannot be allowed to happen.
AWS [cnbc.com] and Azure [techcrunch.com] each do over $20B a year (and growing), adding $1B a year more to either one will not create a monopoly.
I am talking about a cloud infrastructure monopoly when serving and creating a private, sec-cleared cloud infrastructure for the DoD. Once you (the generic "you") get an exclusive contract to create it all, that is, in effect, a monopoly in that space.
Re: (Score:3)
Re: (Score:2)
So sick of people "deciding on a cloud provider." People just don't get it. Being cloud ready means possessing the ability to move to any of them at any time. If you have to decide on a cloud provider you should just stay on-prem until you can put on your big boy pants.
If defense can benefit from "the cloud" for non-sensitive infrastructure then exactly as you say, they should put their effort into defining a spec with detailed QoS and each vendor that wishes to get a piece of the defense cloud pie has to meet the spec, and keep meeting it. But what is this idiocy about classified material in the cloud? It will go horribly wrong. It will. It will. It will.
Re: (Score:2)
Just because their position happens to be right, doesn't mean they actually are right. Remember that they are coming to this position because THEY are not the single-source bidder. If they were the ones on the contract, they would be fighting with everything they have to keep it single-source, because the lock-in they decry in this case is their business model in every other case.
Fuck Oracle. Someone else getting this contract actually IS competition for them, and might force them to learn how to actuall
Re: (Score:2)
As long as none of them are Oracle.
Re: (Score:2)
Re:Oracle seems to be right (can we say that?) (Score:5, Insightful)
This is the first news about Oracle doing something that I think might not be evil that I have seen.
Being right is not the same as being good. Oracle is right, but for reasons of pure self-interest. They got a late start in cloud services, lack scale, and are still sucking hind tit, so they have no hope of getting a big winner-takes-all contract. If they can force the DoD to break it up, they have a good chance of getting some portion today, and even more in coming years.
Objectives (Score:2)
"Objectives:
Acquire a worldwide, highly available, exponentially elastic, secure, resilient cloud computing and storage environment that seamlessly extends from the homefront to the tactical edge."
Maybe it was the 'exponential elasticity', or perhaps the DOD felt like Oracle didn't "extend to the tactical edge" enough.
Re: (Score:2)
I suddenly feel like I want to puke me guts out reading those Objectives.
My prediction is that DoD will go with Azure so they can have Powerpoint Parties and wallow in meaningless slides...but now backed by The Cloud, which MS will convince them will solve their every problem.
Question the Pentagon's use of the cloud (Score:5, Insightful)
We need to drill deeper than simply reporting on Oracle's protest, and the politics behind it.
An independent body of security experts should study the Pentagon's use of the cloud in the first place. Simply by moving to cloud computing, the Pentagon is revealing that they underestimate the cyber espionage capabilities of enemy states, and as in the case of Islamic State or Al Qaeda, stateless enemies.
The same independent body should also study vulnerabilities inherent in military use of the cloud. In an all out war, the enemy first tries to neutralize the command and control infrastructure of their enemy (us). Simply by using the cloud, we are offering the enemy a single neck to chop off, connecting the brain to the body. A secure military force requires so much redundancy, that the enemy has too many necks to chop off to be a feasible strategy.
Re: (Score:2, Insightful)
Simply by using the cloud, we are offering the enemy a single neck to chop off
Your grasp of the subject matter is woefully inadequate. Stop before someone gets injured.
Re: (Score:3)
Broadly speaking, it's a valid concern that we are eager to put all our eggs into as few baskets as possible, and those baskets will have a lot of mono culture in them.
An adversary discovers a way to access some key part of the power infrastructure of a brand Amazon uses and knows of a vulnerability that can deal persistent damage? Poof things could grind to a halt inflicting significant economic damage, and using an attack of a nature that has thus far not justified forceful retaliation in scenarios where
Re: Question the Pentagon's use of the cloud (Score:1)
How is this modded up?
No thanks Oracle (Score:5, Informative)
Bad memories die hard, and your solutions trainwrecked [venturebeat.com] Oregon's healthcare website when other states were able to accomplish more for far less and in a far more timely manner.
Good thing I'm not in congress, I'd find any way I could to prevent you from bidding on a contract that was critical for our national defense.
Just get lost already, and let the companies that know what they're doing get the job done.
I don't trust any of them (Score:2)
I don't trust any of them, do you? Just to be clear: Microsoft getting the whole defence contract would be a disaster. Apple getting the whole defence contract would be a disaster. Oracle getting the whole defence contract would be a disaster. Google getting the whole defence contract would be a disaster. Amazon getting the whole defence contract would be a disaster. Listed in order from most disastrous to... still disastrous? What the hell.
What I want to know is, what is the size of the kickback? To whom w
Re: (Score:3)
Splitting things between different contractors is just going to end up with people pointing fingers at each other.
"It's not out fault, talk to {other provider}"
At least with a single provider you can pin them down.
Re: (Score:2)
Exactly my concern...
It's one thing if the argument is 'you need to award several smaller *independent* providers', so that a total failure of one vendor is isolated and you can keep going, great I wholeheartedly agree.
It's another if the argument is (and I think this is Oracle's hope) 'you need to compose the solution of many providers at different layers', so that a total failure of any of them is guaranteed to knock everything out.
Re: (Score:2)
I am not so sure. There are lots of problems with moving military applications to the cloud. I don't think those can be understated. However one of the biggest problems our government has today is there are to many players in literally every activity it performs. Two men can keep a secret when one of them is dead! Not having this go to all one vendor means multiple parties will have access to the hardware that has tokens and authentication information. Multiple parties will be in a position to observe
Re: (Score:2)
I doubt there is a kickback, it is too open to the GAO and auditing. The DoD is finishing up their first ever audit, that will now become fodder for GAO oversight as it will now be an ongoing audit exercise.
The problem with not going with a mono culture is that congress critters will then use DoD as a punching bag for declaring that it is wasting money attempting to get all the cats herded together. And the extra money it would take to herd those cats will be significant.
Balanced against that is the securit
Re: (Score:2)
I doubt there is a kickback, it is too open to the GAO and auditing.
You are a trusting soul. Look at how Trump flouts the law in broad daylight. There are ways, there are means. Not all payback is in dollars (but most is.)
Re: (Score:2)
Dead company walking. They just don't realize it (Score:2, Insightful)
Dead company walking. They just don't realize it.
Their clients hate Oracle. If they could, they've fire them today.
Oracle has been a bully, especially on cloudy stuff.
And... (Score:3)
If Oracle were awarded the contract instead then Amazon or Microsoft would just sit by idly? Of course not. They would launch lawsuits of their own. This story has nothing to do with what's best for the federal government and everything to do with endless corporate greed.
Obviously Oracle is fighting to prevent a competitor from getting a foot in the door. They want the whole pie for themselves, just like Microsoft and Amazon do.
Oracle warns about vendor lock in. (Score:2)
Re: (Score:2)
any time big gov't contracts are awarded it's typical for non-selected vendors to launch formal protests
See the KC-135 kerfluffle between Boeing, Lockheed and Northrop Grumman.
DoD response (Score:2)
The DoD responded to Oracle:
"I am not altering the deal. Pray I alter it further."
Same thing happened before (Score:2)
Now back to Orac
Re: (Score:2)
Google was unable to develop their own language for Android so they used Java, hoping to tap into that eco system. Which is ironic since Google also berated the Java language.