Pentagon makes case for Return of the JEDI: There's only one cloud biz that can do the job and it starts with an A (or rhymes with loft) The US Department of Defense is pushing back against criticism of its proposed $10bn winner-takes-all cloud mega-deal, dubbed JEDI. The Pentagon this week emitted a flurry of paperwork and presentations including a slide deck [PDF] on the project and an alleged fact sheet [PDF] addressing condemnation of the IT super-contract …
If the military need to fire bullets then they must talk to an oracle rep and he will sell them an upgraded solution that will allow them to use bullets.
On top of that upgraded plan each bullet fired will be charged as an additional cost. Plus expenses.
On a purely technical level, the idea that Oracle could compete on equal terms with AWS on that contract is ludicrous; and given its history of over charging and under delivering, close to immoral. This is the US military infrastructure we're talking about, after all. Of course, given the billions of dollars in computing hardware the NSA, an agency within the DoD, already owns out in Utah, you have to wonder what kind of lame excuse the government could have for outsourcing at all.
"you have to wonder what kind of lame excuse the government could have for outsourcing at all."
The DoD outsource a lot of their data centre management already - moving to AWS or Oracle isn't a dramatic change.
Except for the vendors involved that are losing the existing business.
If IBM and Oracle miss out, it is very likely to kill off Oracles cloud business (it was small anyway and a significant portion is existing US government business) and leave IBM with a headache as they provide a good portion of the existing Data Centre capacity that will be retired.
The government in US exists only to funnel taxpayers money into a few companies which handsomely pay politicians to ensure such system keeps working. More or less happened what Eisenhower warned about.
While I can understand even the DoD can outsource some administrative tasks, I can't understand how they think to be able to support worldwide combat operations through a commercial system, where AWS will try to 'maximize shareholders value', aka profits.
"I can't understand how they think to be able to support worldwide combat operations through a commercial system"
But as you've already pointed out, the raison d'etre of the US military is killing people overseas, whilst also racking up massive profits for defence contractors. They've never had a problem doing it before, so I don't see how AWS is going to be different.
While I can understand even the DoD can outsource some administrative tasks, I can't understand how they think to be able to support worldwide combat operations through a commercial system, where AWS will try to 'maximize shareholders value', aka profits.
Combat operations are not conducted via the unclassified networks, they're planned, communicated about and documented on classified networks that are still DoD operated.
... as a consultant, I'd strongly suggest they hire the correct staff and roll it out internally. When you are dealing with that kind of economy of scale, paying a middle-man for an easily provided service is a mug's game.
Yes, there is historical precidence ... Many of us here are old enough to remember when and why we told the Batch Processing and then the Timesharing Service Bureaus to fuck off.
Yes, there is historical precidence ... Many of us here are old enough to remember when and why we told the Batch Processing and then the Timesharing Service Bureaus to fuck off. .... jake
And as is so typical, with a resulting endless stream of further opportunities and systemic vulnerabilities to exploit and empower quite mischievously, are valuable lessons never learned, jake, for are not managed security service providers (MSSPs) a clone and/or a drone of that which you told to fuck off in the past.
Such suggest an endemic lack of intelligence in humans delivering a rich vein of ignorance and arrogance to be ruthlessly and relentlessly tapped by virtual machines.
The issue is that corporations don't know how to recruit and hire actual security professionals. That's why the snake-oil salesdroids in so-called MSSPs exist ... they have evolved to fill a vacuum. Doesn't mean they know anything about what they are supposedly selling, though, now does it?
\
The rich vein of ignorance and arrogance (AKA "The Board" and other upper management) is being ruthlessly and relentlessly tapped by humans in sales and marketing, not virtual machines.
And all of that both practically and virtually guarantees current SCADA Systems Administrations catastrophic failure, jake ....... leaving Virtual AIMachinery in Effective Command with Alternative Media Control of Ignorant and Undereducated Masses?
And shared as a question because is that not just a logical development of the way things presently are commanded and controlled, albeit extremely badly in the service of vested interests/investing principals?
Alien8ing Operations ‽ ..... on Heavenly AIMissions ‽ .
:-) amfM Hailing Futures and Derivatives Marketeers .... Quantum Communications Pioneers with Virtually Private Pirate Intellectual Property Enterprise Entrepreneurs ...... and urPresent Hosting Program advising One and All of a Program Change with Fundamental Systems ReProgramming with Novel Input to Nobly Output in Scenes from these New Fangled and Entangling Virtually Realised Productions.
And most definitely QuITe Titanic Studio Type Stuff for Advanced IntelAIgent Intellectual Property Programming for Realisation with AI and IT Providing All Worlds a View of Lives Loved and Lived Elsewhere Somewhere ...... via the simple expedient of Trial Run Undeniable Media Presence.
Quantum Communications Leap Time, MOD. In with All or Out with Nothing?
cc Palace Barracks
And if the company said "our current spend is U$1.2bn/year and is predicted to rise to $2+bn/year within the next 10 years, how can you reduce our spend and do so as quickly as possible?"
This is about trying to control IT vendor spending. If this is delay by a year, the $1.2bn is likely to be $1.5bn. Instead of paying every vendor that meets the FedRAMP requirements for DoD cloud, they will look to reduce the number of vendors (and duplication) significantly. They are culling off the middle-men.
Oracle and their associates are trying to slow JEDI down because once it is operational, DoD dollars will go to JEDI instead of existing data centre/cloud providers. Oracle, IBM and many other DoD providers will lose out as JEDI results in the closure of some of the legacy Data Centres in the DoD's 300+ DC fleet.
Oracle and IBM's response was that they would take ~2 years and $200m-$300m to purchase the land and build the infrastructure and then they would work on the rest. And charge $1bn/year. I know its a little more subtle than that, but it's a likely outcome.
"how can you reduce our spend and do so as quickly as possible?"
Same way I always do it ... First we get rid of all the middle management who exist to do nothing but attend meetings and trade Power Point presentations, and the flunkies required for their care and feeding. That usually cuts staff by around 15% in medium to large corporations.
First you're starting from scratch and now you're firing middle management? The DoD's problem is primarily vendors charging significant amounts for not much. In cloud, the primary issue is vendor sprawl where each vendor has to meet significant requirements before hosting services on what is effectively a private network with heavily controlled access to the public Internet. DoD vendor's pass ALL costs to the DoD. JEDI does to data centre vendors what you are suggesting should happen to middle management.
While Oracle are desperate to make this look like anything but what it is, the existing landscape is:
- a large existing IT footprint, largely managed and maintained by third parties.
- the third parties are used to quoting for gold plated solutions and having them approved, often against DoD IT's recommendations. Most of the gold plating is around the sales peoples bonuses and the vendor mark ups, with any costs being charged to the DoD. The vendors don't want to lose this because outside of the DoD they won't get anywhere near the commercial returns on what are often legacy DC's
- the DoD have been attempting to work with vendors to achieve consolidation and cut spiralling costs since 2005 under 3 administrations and have been struggling to achieve saving.
- Oracle and IBM's bids for JEDI were uncompetitive on price and time frame because they included requirements for the DoD to fund the purchase of land and/or facilities for new data centres as well as any time to get the facilities ready. All up this was expected to make a difference of $200m-$300m and 2 years to their bids, leaving the decision down to AWS and Azure.
- the high level strategy is that Office365 apps end up in Azure and other backend apps will be moved from legacy data centres to either a state level facility (if latency is an issue) or an AWS facility. There will continue to be exceptions for special project facilities (i.e. the F35 DC or Google's assistance with satellite image processing), but the intention is for a significant number of legacy facilities (and associated support/maintenance) to be removed from the DoD infrastructure.
"When you are dealing with that kind of economy of scale"
What scale? With only $10Bn to spend over 10 years you wouldn't even get a discount with Intel as you'd be a tiny little customer next to Amazon, Microsoft, Google & Facebook. You wouldn't be able to afford custom hardware, firmware or CPUs which would make everything else more expensive. Even the cooling at Amazon and MS is more efficient because it runs warmer thanks to custom Intel chips which are set to higher thermal limits.
And that's all ignoring that AMG&F have hired every single* infrastructure, networking, security and app specialist worth hiring, so there is pretty much nobody left you could hire who could build a reliable cloud without blowing the whole budget on wages, bonuses and security clearances.
*OK, I'll accept there are probably a few hold outs, but certainly most of the good ones will have been assimilated. Certainly all of IBM's good people took their redundancy money and jumped to real cloud vendors a while back.
"hire the correct staff and roll it out internally"
In theory, yes. Unfortunately government projects tend to become a hierarchy of gradually declining competence. The implementation comes down to people randomly poking at the system until it somewhat matches the big checklist of inflated and contradictory requirements. There's then an infinite series of costly feature realignments.
The $10 billion budget alone is going to scare away anyone who likes a project to be organized and done right. Ironically, it could be more successful and accomplish more if it had much smaller goals.
If we put aside Trumps presidential style, the companies lobbying for political pressure to change the outcome would likely have achieved the current result. The final result will be interesting regardless of outcome:
a) JEDI stays with AWS. Good for Amazon, bad for Oracle and IBM, bad for other smaller DoD data centre providers
b) JEDI is given to Azure. The DoD's cloud strategy is basically Azure for Office365 functionality and AWS for all other generic requirements - putting everything into Azure is a big win for Microsoft but likely a large loss for everyone else unless one of Oracle/IBM/AWS come up with an Office365 replacement. Azure has a way to go to meet AWS's physical locations, so this would likely drop resilience against attempts at disrupting DoD infrastructure.
c) status quo. DoD cloud spending likely continues to spiral as the DoD tries to consolidate legacy data centres into multiple cloud providers. The DoD ends up with 70+ cloud providers, most of which barely meet the requirements for resiliance but further political lobbying makes
d) a compromise is reached between AWS/Azure/IBM/Oracle leading to more than one vendor being "awarded" the contract to avoid slowing implementation with lengthy legal battles. Given the minimum spend of $1m, maybe spending $11bn (AWS plus IBM/Oracles setup fees and minimal on-going spend) on a $10bn contract is the cheapest option?
e) something else. I'm sure there are a myriad of other options.
There is a lot of DoD spending at stake here, with the DoD spending almost $20/bn a year on IT, although this is likely to be on the low side once special projects and other methods of getting IT systems in-place are included.
Any multi-cloud approach has some significant drawbacks;
1. Legal battles when there is a problem/delay/breach will make the current delays and court action seem like a pleasant day out at the beach
2. Maintaining security standards will become a nightmare and be more likely to lead to a breach
3. The overhead of managing across multiple providers will negate any savings and have the opposite effect of increasing cost through technical complexity and also decision making (which cloud for system X) along with the associated on-going lobbying (see point 1)
4. They're just the problems I can think of in 2 minutes - think of what problems Oracle and IBM could come up with given a little more time!
While I agree that multi-cloud makes things harder, JEDI is about scaling back from 50+ vendors to maybe 20-30.
The points you raise are valid, but the DoD is coming from the opposite direction - they have a complex infrastructure that they have spent a significant amount of time securing (hence FedRAMP requirements to provide access to the cloud - while these were initially a possible way of limiting cloud provision to fewer vendors, it didn't work out that way, you just had replicated functionality).
As far as I am aware, security standards have not been a major issue as non-compliance means the DoD can disconnect you and move services away from your hosting facilities. However they have been a major driver for cost increases.
Point 3 is the fundamental reason for JEDI and point 4 is actively in-play as a consortium of vendors headed by Oracle is seeking to disrupt JEDI (i.e. "it will be outsourcing to a single vendor", "the bidding process was unfair", "lack of resilience", "preventing innovation") - almost all of the points that the vendors used to try and keep AWS/Azure out of bidding in the first place.
You forgot this option: Oracle don't need to win.
JEDI is around $10B of business. Let's say Oracle use $10m on a lobbying effort, and because of the fuss they kick up win 10% of that business. That's a massive ROI and Oracle will cry about losing all the way to the bank.
"You forgot this option: Oracle don't need to win."
This falls somewhere between status quo and delay and is probably the most likely outcome of political interference - Oracle are fighting for a graceful exit from GovCloud to avoid providing bad news to shareholders or at least providing it as part of losing JEDI when the reality is slightly different. i.e. without GovCloud, Oracle's cloud offering is likely to be wound down as uncompetitive.
"the DoD said in explaining its one-provider specification. "It is important for a warfighter in Afghanistan to access the same information as an analyst in Washington, DC or a service member training in California.""
So what happens when that ONE cloud provider has an issue in that "region"?
"the DoD said in explaining its one-provider specification. "It is important for a warfighter in Afghanistan to access the same information as an analyst in Washington, DC or a service member training in California."" .... steviebuk
Who provides the Afghan warfighter intel on all who choose to be mortal enemy rather than eternal friend? Is it feared in the West when provided by Allah? ...... with Supplies Proven Virtually Almighty and Practically Inexhaustible.
This type of contracts seem nice and an efficient use of federal budgets.
For the civilian market it implies that the big cloud providers become military targets as well.
In case of war, China may start targeting data center locations and head quarters of Amazon and MS first, as part of a strategy to reduce the capabilities of the US military.
It is questionable if the big bit barns in VA are blast and EMP proof.
A Pearl Harbor like strike addressing Amazon and MS may yield interesting outcomes if done well, not taking into account the unpleasant collateral damage for the people living nearby those facilities.
1) Don't forget that the target in war is not the military, but the policy makers.
2) Inflicting (preferabbly) non-fatal pain on civilian populations in a popular government is the most efficient way to get policy makers to back off in a war.
3) Open borders allow the preposition of a substantial amount of munitions anywhere in the US.
Conclusion?
Add civilian data centers to the list of critical infrastructure that goes "poof" as soon as a war breaks out with China. Much easier to target than the military ones, and they have far greater effect on policy makers.
Add civilian data centers to the list of critical infrastructure that goes "poof" as soon as a war breaks out with China. Much easier to target than the military ones, and they have far greater effect on policy makers. ..... Claptrap314
Here is evidence of the Moron Targeting of a civilian date centre and the Rupert Murdoch empire, News Corporation. ....... Satire? Serious? Sunday Times prints column asking to ‘give war a chance’ & attack CHINA.
Nice one, Rod ....... not.
What a plonker.
C'mon, amfM, you know better than linking to this kind of thing ... Liddle is an attention whore, no better than your average loud-mouthed 14 year old on 4chan (or 16 year old on reddit). He probably doesn't believe a word that he types as long as he gets listened to (and his paycheck). Don't give him what he craves, bullshit artists like him aren't worth wasting electrons on.
Whilst we can easily most probably agree on all of that, jake, the post was more of a lament on the parlous state of News Corporations harbouring and giving succour to the purveyors of abhorrent shenanigans, and for which the parent host can be held jointly and severally liable.
Who'd have a thunk it? Attention whores infecting and injecting the Times with poison pen lettering?
I'm pretty dang certain that good ol' Sam Clemens observed your exact lament back in the 1800s, and he was far from the first. Letters to the Editor authors writing drivel and attempting to affect TheGreatUnwashed has a long and storied past (thus the comedy staple). Some of these twitiots have a gift of the gab and get hired full time. Somehow, civilization has muddled through.
I remain, Unperturbed in Sunny Sonoma
AWS is global, so any mythical attack on AWS would involve a global attack on dozens of data centers around the world. So, a military attack is out of the question, leaving things to network centric attacks, which are a daily event anyway.
"AWS is global"
But GovCloud isn't.
Having said that, it would appear to be widely dispersed. If you can smash Washington DC/Virginia/Maryland you've likely taken out a significant chunk of GovCloud with or without JEDI.
To do that requires the nuclear option at which point most computer systems will struggle to survive.
AWS/Azure GovCloud facilities would likely be at 50%-67% availability at that point as they have presence in western and southern US locations too.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
