166 comments

[ 2.7 ms ] story [ 164 ms ] thread
The Department of Defense’s (DoD’s) lack of a coordinated enterprise-level approach to cloud infrastructure and platforms prevents warfighters and leaders from making critical data driven decisions at “mission-speed”

From the Statement of Objectives [1]. Though it's probably more about the "Online marketplace offering" mentioned. I can see it now, bunch of office workers deploying some PHP software to their JEDI cyber cloud.

1: https://www.fbo.gov/utils/view?id=9283ca4f1d895893346235548a...

> the "Online marketplace offering" mentioned. I can see it now,

Generals who bought this item also bought...

"Now with multi-colored tracers!"
When I toured the Las Vegas Switch data center a month ago, I asked what all the AR-15s were for. The response was that it is required by their all ex-military security team as a requirement to host highly classified government servers.
HA! those government requirements
It reminds me of this desk from silicon valley: https://regmedia.co.uk/2016/05/09/silicon-valley-ep3-table.j...
Is that from the TV show, or the actual Silicon Valley? I didn't watch the show, could someone explain the reference?
It's from the show. At that point in the show, the main company in the show had packaged their compression technology into a rack mountable device, and part of their service agreement with customers was a requirement for 24 hour support. The guy in the photo with the pony tail is a datacenter employee who is showing the main cast where their support personnel will have to sit. It's basically a waste of time and space to have someone at that desk 24/7, but it is part of the contract that their customers required.
Reminds me of Raymond Hettinger’s talk where he basically goes, “Government regulations may seem weird but they usually have some sort of disaster behind them.”
"Regulations are written in blood."
I'd imagine you mis-understood the statement...

I previously used to do "Physical Security Assessments" of Classified environments. An armed guard was never a requirement... The data center should be following physical security requirements listed in AR 380-5 [0]

As for the all ex-military bit, that's just b/c veterans are the only people with security clearances and little to no professional skills. The DoD will give an 18 y/o infantryman a security clearance just b/c he asks. Good luck getting one from industry or any other government agency if you haven't demonstrated a 100% need for one in a professional capacity.

[0] https://armypubs.army.mil/ProductMaps/PubForm/AR_Details.asp...

> 18 y/o infantryman a security clearance just b/c he asks.

Not true at all. Most enlisted infantry don’t have a security clearance. An standard infantry platoon officer typically has only a Secret clearance — only getting Top Secret in special cases and with special units. An 18 year old enlisted Ranger would have a clearance, but some grunt in the 3th Infantry Division almost never has a clearance unless there is a specific mission need.

The standard for clearances is based on the need to access classified information. An industry or professional partner working in those areas has a higher risk profile because they are typically accessing or building classified systems — which is a far greater risk than the access afforded to a typical infantry enlisted soldier.

clearances for non-military working on DoD projects were, until 2018 handled by OPM which has a significant backlog. So that made OPM clearances much slower. However in 2018, the law was changed to allow DoD to bypass OPM for clearances.

As far as ex-military with little or no professional skills, that’s a bit insulting and heavily generalizing. It really depends on their role. If they are physical security, then “professional” skills (as you might define it) aren’t relevant. Their job is physical security. As far as other ex-military with no professional skills, that’s ridiculous. Many ex military have highly developed professional and technical skills.

But nobody gets a clearance “because he asks.” An 18 year old kid in the infantry doesn’t get to ask for anything, let alone a clearance.

Also, an 18 year old has considerably less background to check than a 35 year old software engineer. I had a clearance from State that took 18 months to process because I had lived in 5 different countries, including countries that weren’t considered “friendly.” But an 18 year old who lived his whole life in the same town — those clearances are issued very quickly.

And you should have a 100% need to get a clearance. I am not sure why that would even be worthy of debate.

When I was in my early 20s, I picked up a confidential clearance in a very short period of time. IIRC it only took a couple of days.

The need was very well defined. I required escorted access into a missile lab to service their network equipment. My background was also quite limited. I'd never traveled much at the time, lived in a couple of states.

Confidential clearance does not seem stringent. You can avoid reinvestigation for 15 years, while secret needs to be reinvestigated every 10 years.

I've generally heard that secret level work is pretty boring, so I doubt confidential would be better.

You're right about secret level work being boring, it's normally related to stuff like the exact time and location of a military flight or convoy, or other things that can impact people at a smaller level, but aren't really dangerous to the nation as a whole if they get leaked. I think I only ever saw something marked as confidential one time while I was in the military; general information was normally unclassified and anything that was deemed sensitive was bumped up to secret because it was easy to get for anybody that had a need-to-know.
We're starting the get into the finer details of the security clearance process, which I don't really want to do. But I'll do a high level response.

I said what I said b/c I've personally been tasked with getting an entire infantry battalion (~900 infantryman + the support elements) security clearances for a deployment. The battalion had no need for security clearances, but the DoD had approved all unit expenditures related to the deployment, so the BC requested it b/c a few Joe's asked for it so they could improve their job outlooks post-military.

This comes down to the difference between needs for a security clearance and needs to access classified material. Just about anyone can technically get a security clearance investigation. Having completed a security clearance investigation doesn't actually award you a clearance, it awards you the eligibility for a clearance. Example, if you get your clearance in 1st Battalion, after you pass the investigation you get read on and sign an NDA (SF-312), the Security Manager accessing a system called JPAS and with the click of a button "turns your clearance on" (for lack of a better phrase), that awards you a clearance. Now lets say you transfer from 1st battalion to 2nd battalion, when leaving 1st battalion you sign another NDA (SF-312) and acknowledge that you understand what you did with that unit that was classified (often times you did nothing...), the security manager then goes back into JPAS and turns your clearance off. When you arrive at 2nd battalion, their security manager has you sign a new NDA (SF-312) and goes into JPAS to turn your clearance on.

So yes, and entire infantry battalion can have security clearances without serving in any classified environment.

I didn't say everyone in the military has little to no professional skills, so please don't put those words in my mouth. The military is just the only organization that gives clearances to those with little or no professional skills. Justice/State/industry sure as hell don't.

You misunderstood a slightly ambiguous sentence. The parent claim was that of you need low skilled people with clearance, your main candkdoo pool is full of vets. Not that if you look at vets, they are likely low skilled with clearance.
> veterans are the only people with security clearances and little to no professional skills.

That is a horribly bigoted statement, the type of which wouldn’t be tolerated on HN if it were directed at almost any other group. You ought to be ashamed!

The comment clearly wasn't about all veterans.

In a case like this, the following site guideline should kick in: "Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize."

https://news.ycombinator.com/newsguidelines.html

Are you sure that wasn't Switch in Las Vegas? They've had armed guards for over a decade. My colleagues would comment positively about this fact, "they even have armed guards" as if that's a good thing.

No service that I've ever run has ever required armed guards and certainly don't ever want to have my services colocated with other services that require armed guards.

Longer than a decade. I saw the AR-15s in 2006-ish. They also made a big deal that they are allowed to seize diesel fuel from all the surrounding gas stations in a real power emergency.

I just think it’s more marketing campaign on their part, honestly.

That was my take as well.

I remember the diesel fuel thing, too. Dumb.

It will be interesting to see how this plays out considering the huge controversy around project maven & google. I fully expect the HN moral outrage machine to kick into overdrive when this is signed.

I would suspect that the major providers would be better off to work with a third party on this, rather than run it themselves. E.g. had Project Maven been done through some random contractor who used GCP just like any other customer does, I dont imagine it would have got the same level of hysteria that it did with Google being the main contractor.

Amazon already provides the IC's main cloud. I'd imagine they're expected to win this one as well.
(comment deleted)
Do we want to speed up war? Some dev accidentally merges the "launch the nukes" PR, and nuclear apocalypse commences.
Even if the US doesn’t do this, you can bet some other countries like China will have no scrupples with investing in capabilities. US military dominance shouldn’t be taken for granted. I would expect China to eventually surpass the US given the people advantage, more top down decision making across the board (protesting unpopular decisions is less tolerated), and China’s rapid progress in technology.

I don’t think this is as simple of a decision as some tend to make it out to be. Technology tends to be one of the factors that reshape geopolitical spheres.

> US military dominance shouldn’t be taken for granted.

Should it be taken as positive on the whole, though? Dominance is a reality that we live with, and one that the US certainly benefits in some ways from. However, there are another 7b (or so) people on the planet who suffer the consequences of our military dominion (their backyards), national interests (their natural resources), etc...

> Technology tends to be one of the factors that reshape geopolitical spheres.

there's a bias on HN towards blaming "the tech" for all "the things"... how about: social unrest, resource extraction, and empire building? these are much less tech-centric, yet real and studied causes for geopolitical reshapement.

I bet if you looked at all the possible universes following World War 2, the Marshall Plan and ensuing US hegemony sorts fairly high even when accounting for the entire world.
"sorts fairly high" from whose perspective? Certainly that worked out well for the US, Allies, etc...

what about the Palestinians? the Syrians? The Yememi? The Chileans? The Colombians? The Vietnamese? The North Koreans?

take a look at a history of US meddling in the world... and consider sort of world we might have it the US military and its industry (and those of its allies) hadn't nosed it's way into all corners of the world pushing for hegemony of "Democracy", the US dollar, and their economic interests.

I said sorts fairly high even when accounting for the entire world. to begin with, I didn't edit that in or anything.

In a whole bunch of possible outcomes you have Soviet failure everywhere.

yes, there are lots of instances of technological progress, improved medicine, etc...

> In a whole bunch of possible outcomes

in a whole bunch of actual outcomes, you have US failure everywhere

“We are the ally of the US not because they are powerful, but because we share their values. I am not surprised by anti-Americanism; but it is a foolish indulgence. For all their faults and all nations have them, the US are a force for good; they have liberal and democratic traditions of which any nation can be proud. I sometimes think it is a good rule of thumb to ask of a country: are people trying to get into it or out of it? It’s not a bad guide to what sort of country it is.”
My comment on military dominance was a neutral one - I was making no judgment on it being a good or bad thing. However, if you are an American, you almost certainly would rather have it where your country is in command versus the other possibility (Note: I am a first generation American who has served in the Marine Corps as an infantryman).

My comment on technology is not just confined to electronics - there is a long history of military technology affecting battles, and perceived advantages/disadvantages of countries, which results in decisions being made according to that calculus. One small slice of that history is the introduction of guns to Native Americans. I purposefully did not say it was the only cause, and it is a complete misreading of my comment to state otherwise.

to your point... there are ~350m "Americans" and all the rest of +7b humans aren't. Whatever the desire of Americans in their country, democracy would suggest that +7b people want something that looks like their country in a dominant position.

> complete misreading of my comment to state otherwise.

great... I didn't say otherwise.

There is a lot of money being dumped into war capabilities, and as an American I would rather it go into making lives directly better rather then killing for unclear overall benefits.
(comment deleted)
"Mr. President, we must not allow a mine shaft gap!"

~ General 'Buck' Turgidson

#DrStrangelove

> more top down decision making across the board (protesting unpopular decisions is less tolerated)

Maybe they'll get pretty good at playing catch-up with this strategy. Also note that this suppresses intellectual freedom. US will have "people advantage" for the foreseeable future because of diversity and education. For e.g. A chance meeting between myself and an elementary school principal for a school in rural Oklahoma led to a (free) "Intro to South Indian Classical Music" Skype session with some kids and my sister who lives in India! This scenario is possible in China as well, but more probable in America now and the foreseeable future.

The point is that exposure to this kind of diversity sparks creativity. Google "chinese education rural vs. urban" to see the gap. This gap is not so great in America. So coupled with diversity and better education, US students are much better placed to forge the future!

What's the point of being "ahead of china" in terms of war? Isn't one of the great benefits of MAD that it provides an upper bound on needed destructive capabilities, beyond which you are just wasting time?
The point in being ahead of China is that MAD is the upper bound, but there is still a lot of destructive room to grow between current capabilities and nuclear apocalypse. If a country has nukes but a weak conventional military then you can be sure that the world will do everything it can to control your nukes, and even then the weaker country knows that if it uses nukes, it'll be wiped off the Earth in retaliation.

A good example country is North Korea; they have nuclear weapons, but there conventional military is so weak compared to their neighbors that the only thing nukes get them is food aid.

There's also the matter of limited warfare and not being able to use nukes for every defensive purpose. A nukes-only approach to national security means that you nuke over the slightest provocation, which will earn you short-term victory at the cost of international condemnation and likely intervention. No one likes a nuclear powder keg.

For a practical example, let's say we never bothered to invest in point defense for our destroyers and they were completely vulnerable to Chinese anti-ship missiles. One day China sinks a destroyer as a show of force. Should we nuke Beijing? Or would it be better to have a strong Navy/Air force in place to respond conventionally by taking out the missile launchers?

> Do we want to speed up war?

If you think there's any possibility that your country might have to fight any kind of war, including a war of defensive national survival, then you better hope they're able to do it quickly when it happens.

When was the last time the US fought a defensive war? I think we can argue that WWII, was ultimately a matter of national survival at some remove, but it certainly wasn’t defensive. I think it would have to be 1916 when Pancho Villa attacked New Mexico, and that wasn’t exactly unprovoked either.

edit Why is this comment automatically closing?

War on Drugs.

War on Poverty.

Afghanistan.

"war on poverty"

Here we go again. Using the word "war" for all sorts of things that have nothing to do with actual war, even in a discussion about actual war.

(comment deleted)
(comment deleted)
For instance, it's worth noting that [not Chile] Costa Rica doesn't have (and hasn't had) a standing army, but has successfully fended off multiple attacks from neighboring countries.

The US is the most heavily armed country in the world, and it's domestic population is not unaccounted for in terms of armaments.

edit - I realize it's considered uncouth to have a meta-convo about downvotes, but why the f* is this comment downvoted?

edit2 - stupid dummy... I wrote Chile & meant Costa Rica

This is just wrong.

Chile has a substantial defense capability, allied with the US. We host their cadets at our military academies, and ours at theirs. Maybe you're thinking of a different country.

OMG, you're totally correct!

I wrote Chile & meant Costa Rica. Thanks :)

Right, mostly cause the USN/USCG has a cooperation agreement with them for maritime defense.

There's no such thing as a country with no defense forces. You either have your own forces or are aligned with a major power (US/RUS/CHN).

Costa Rica here.

Not really much to plunder my country for, and if they wanted to, they already did (enclaves about a century ago).

Armed conflict is only necessary if you don't give people what they want.

Sure, carry a big stick and such. But that doesn't mean it is without problems.
We often let the ideals of perfection get in the way of progress. Wouldn't it be great if we could live in a world free of nukes, free of war, free of the massive amount we spend on these sort of things?

Unfortunately, we're not there, and it's probably not a stable equilibrium. Instead, history has shown that complacency does not lead to peace.

So what to do? I think the correct answer is to match other countries capabilities, but always always always take a "peace first" approach to everything. It's far from perfect, but it does lead to progress. As such, such things as analyzing drone footage with AI, datacenter contracts, satellite bandwidth contracts, etc, are a necessary part of the peace process.

This comment doesn’t fit with the pacifism at all costs mentality of many, however, it’s by far the most erudite point I’ve read in these parts in a long time. Peace is achieved only through strength. ISIS for example doesn’t care about our lofty ideals of peace. This is/was a group that burned prisoners alive and gang raped kidnapped women. They don’t respect treaties or pontification on civility. If we as a free society are to have peace, we must be prepared for war.
Exporting democracy with bombs has not worked well so far, I guess (or hope) governments learnt that. Terrible mistakes has been made, justifying war at all cost to satisfy people for election, or other wrong motivations.

You can't just destroy a regime, replace that with another democratic one and expect that a State of peace will grow. This just doesn't work. People from different areas of the world are used to different way of dealing with power, for cultural and religion issues. This is especially true with stabilizing an extremist Muslim regime. You need to cut their extreme belief, destroy their motivation, give them other chances. Examples: Gaza, Paris and its suburbs, etc.

Most of the time, in those contexts, war doesn't resolve anything.

They never planned to export democracy after all.
The United States is by far the greatest promoter of war on the planet. There is absolutely no goal of peace within the US Military Industrial Complex.
This logic could work for Switzerland. What if your country started numerous wars in the past?
My guess is Oracle Cloud will make the most aggressive bid. Getting in bed with the government and negotiating long term contracts is how they compete.
Yep, and it's not just quantity over quality with them, but also legal over quality... Business-wise a sound strategy, but morally bankrupt and technically offensive.
Have you used the Oracle Cloud Infrastructure product, formerly Bare Metal Compute? It's not a low quality product, by any means. I'm not sure I understand your criticism, unless the point was to be intentionally nebulous.
>My guess is Oracle Cloud will make the most aggressive bid

Not sure , it's also about sells arguments.

Governement often runs on Oracle because it was the most popular RDBMS at the time with enterprise support. So this is a massive argument technically speaking to convince customer : "We'll migrate your oracle DB and infrastructure for you , we have the tooling for that".

> Governement often runs on Oracle because

...outright corruption, though that coming brilliantly to light in the 00’s in California is also why comparatively little in California State government does any more. (Also why the trend toward IT centralization in the State was stopped dead in favor of decentralized decision making with central oversight.)

They'll underbid, overpromise then under deliver, then contest the lawsuits, then bilk someone else.
They'll underbid, overpromise and still lose to AWS. Oracle is amateur league in cloud computing.
It doesn't matter what they can do. It only matters what they can convince you they can do.
It's how everyone specializing in government contracting works, and (except for who they get in bed with), basically the same in enterprise contracting more generally.

IBM, HP, and even Microsoft, among cloud vendors, are all experienced in this game, even if Amazon and Google might be less so. (Though both have five at least some, and aren't neophytes.)

AWS already has 2 air gapped regions solely dedicated for the US government. With that experience and all of their services they can offer, my guess is they will win.

Edit: the sub-linked article is much better [0]. Here is a good quote from it: "In the RFP, officials noted the contract will be competitively awarded, but Amazon Web Services is widely considered the front-runner for the contract because of its existing contract with the CIA and ability to host secret and top-secret data. According to the RFP, the winning company must meet the government’s rigorous standards to host classified data within 180 days of the contract award, and meet standards to host Top Secret/Sensitive Compartmented Information within 270 days. Currently, Amazon Web Services is the only cloud service provider that meets those standards, though officials from Microsoft and IBM say they’re close to meeting them."

[0]: https://www.defenseone.com/technology/2018/07/pentagon-accep...

The scale of oracle cloud is laughable compared to its competitors.

http://www.platformonomics.com/2018/05/follow-the-capex-sepa...

In terms of value to the taxpayer I think AWS's offering would be head and shoulders above anything else. There would be no quicker way to be fired at amazon than if evidence came out that you were rorting or delivering an inferior product to a customer, behavior thats almost assumed with other government contractors.

Interesting. So, if oracle wins, this will be the only customer they care about with its small cloud? Meaning à win on this bid, will likely make oracle cloud less attractive than before, to anyone on the outside?
Using throwaway account.

This is all Oracle’s doing. AWS was originally awarded this contract but Safra Catz (co-CEO Oracle) is on Trump advisory committee. She planted the seeds that Bezos can’t be trusted due to his ownership of Washington Post and what that publication has been saying about Trump.

Fully expect Oracle to make a huge run at this business. And don’t forget, Oracle was started as a company from an initial contract with the CIA. There’s not a single tech company more aligned with the US government than Oracle.

https://www.bizjournals.com/sanjose/news/2018/04/16/pentagon...

On the other hand, when the IC wanted a cloud the only two real competitors were IBM, Amazon and the NSA. IBM protested the award but Amazon ultimately won.

If the requirements for this are anything like the ICs then Oracle doesn't offer enough to truly compete. It'll be either Google or Amazon in the end with Microsoft winning a side contract to provide Exchange, AD, and Office365.

To be fair, when the IC bought their own AWS cloud, Oracle had just unveiled their Bare Metal Compute product. Only in the last year or two have they begin to offer virtualized resources in their cloud. AFAIK, the virtualized offering runs on top of their BMC foundation and the product has been renamed to Oracle Cloud Infrastructure.
They still only have a couple of data centers.

"Oracle Cloud" is vaporware.

More than a couple. Not as many as amazon.

Source: I work for Oracle Cloud Infrastructure

You're absolutely right. The last time I checked it was Phoenix and Ashburn. Quite a few more, now:

https://cloud.oracle.com/data-regions

I'm sorry that you work for Oracle.

Oracle actually has dozens of data centers all over the world. How many "public" cloud users they have is totally obfuscated and in my opinion is probably an outright fraud. Oracle, please don't sue me for slander. I said probably, not definitely.

In reality, I suspect most of these data centers are just private co-locations.

I haven't read the RFP but I doubt the number of data centers will matter. Odds are that these would be placed on federal property, such as certain military bases, and not be a part of any preexisting commercial infrastructure.
BMC isn't enough though. Amazon didn't win simply because they had EC2.
Another thing to consider is how Oracle structures it’s ULA (unlimited license agreements).

Oracle explicitly does not allow companies to cerify their software on AWS.

So just think about how much Oracle software the pentagon runs.

It’s hugely compelling to not go with AWS because Oracle won’t cerify nor grant you a license if your certify your ULA on AWS.

Financially alone might dictate going with Oracle Cloud over AWS. Regardless of which cloud service is technically better.

https://www.itassetmanagement.net/2018/06/12/are-you-sure-yo...

It seems likes there’s a lot of good reasons for oracle to get a chance apart from the wapo stuff.
Giving competiomtors chances is great. But we need more competition than Kodos and Kang. The government should be promoting wide competition and open protocols and interop to avoid lockin.
Or given their budget, why not operate their own cloud? Why not build a tech/infrastructure team and run their own datacenter/cloud? You would think something so important as data and the cloud would be something the pentagon wants to keep in-house.
Fewer kickbacks if you operate inside the government.
Because it would cost the government $100B to do it themselves.
You need serious talent to accomplish projects at that scale, and most of that talent does not (want to) work for the government.
It's unfortunate (and likely economically sensible) that most cloud providers dont offer Power hardware. Running Oracle on Power systems is probably the only reason anyone still runs AIX these days, nd many orgs haven't gone to the cloud because of huge Power/AIX footprints. Even without a ULA in play, licensing on Power systems is significantly more attractive than running on Intel when it comes to Oracle.
Is the software to create a public cloud on Power chips a thing? Can Xen run on power? Or OpenStack?
I know OpenStack/KVM is up and running on Power, not sure about Xen.
Are you sure it isn't because key Google and Amazon employees have threatened to quit their jobs if either of these companies helps their own government? I mean, if I were USG, I would be running away from Google.
"Government" is is not the same as "weapons". Your comment only applies narrowly to the DoD.
This is a $10b contract for the Pentagon, not just generic "government." It is a contract that is guaranteed to involve weapons, war and killing people in numerous ways.

The Pentagon would be crazy to do business with Google after their employee revolt. They can't afford that kind of elevated ideological-based risk and instability.

I am going to give you a down vote because the conspiracy theory is completely at odds with the absurd bureaucratic complexity that is the federal contract bid process. Congress, since we are ultimately talking about money, has far more say in this than does the president regardless of the agency.

> There’s not a single tech company more aligned with the US government than Oracle.

I am going to have completely disagree with that too. If you have ever worked IT for the federal government, especially the military, the most absolutely clear answer there is Microsoft.

Also, as somebody else pointed out, thousands of Google employees recently wrote a protest letter to their CEO threatening a massive boycott over an existing tiny DOD contract if Google didn't pull the plug. That being said why would I consider them for another DOD contract that is thousands of times larger?

The issue wasn't with the DoD portion of it, it was the drone part. I'm sure Google employees would be fine providing Google Cloud and Docs functionality to the DoD bureaucracy.
Irrelevant. Not matter how well intended those employees' motivations it likely cost their employer a substantial piece of this enormous profit pie.

Google employees don't get to choose their customers. They certainly don't get to choose candidacy in a federal bidding cycle. If there is confusion on this matter then perhaps you are not a sells person or a business stake holder.

Although this is amusing theory, it's non-sense. Google Cloud already has government users. Google employees protested an internal project - a thing that they were working on or being told to work on - not a customer. Regardless of how media reported on it.
It was a project under contract to a customer. That is not an internal project. It would be risky and foolish for that exact same customer to gamble their next contract under those conditions.
First of all...

1. They didn't cancel the contract. Google pledged to not renew the contract in 2019, and established a "code of ethics" for AI usage, in response to employee outcry.

2. You're implying that this is a game of hurt feewings and it isn't. It's business. The government agencies have just as much to learn from PR backlash against themselves as Google does, because they hire from the public.

I never said a contract was cancelled. It is, though, a termination of business.

> You're implying that this is a game of hurt feewings and it isn't.

I also never said or implied that. I have worked all over the country and a few times out of the country. Only in the bay area and Seattle are people so social engrossed with intermingling their personal politics with their work. Most other locations, both inside and outside the US, would consider this ethically collusive. Whether or not that hurts peoples' "feewings" (as you put it) is up to the individuals involved.

>It is, though, a termination of business.

A promise to not renew a contract is not even the same as not renewing the contract. For all we know, Google might renew it anyways. Either way, Google is doing what they were paid to do for the period they were paid to do it. The DoD has no reason to have sour grapes with that. Other government agencies absolutely have no reason to.

>I also never said or implied that. I have worked all over the country and a few times out of the country. Only in the bay area and Seattle are people so social engrossed with intermingling their personal politics with their work. Most other locations, both inside and outside the US, would consider this ethically collusive. Whether or not that hurts peoples' "feewings" (as you put it) is up to the individuals involved.

I don't know what you're getting at, I've worked across the country as well (though not out of it.) The consensus among workers is different because the power dynamics are different. When I was working in the midwest, tech jobs were rare and most people did not want to put their jobs in jeopardy.

Employees at Google just feel that they have job security, because they do. Not just those in the Bay Area, but certainly those in Pittsburgh, or New York City, as well. The proof is in the pudding: to the best of my knowledge, Google employees were not fired for protesting Project Maven.

But of course people in the Bay Area feel especially safe in their endeavors. Again, they are. If a Google employee gets fired for protesting, plenty of other bay area companies will be eager to sweep them up. Some will even throw in a sweet signing bonus to welcome them in.

I don't have strong opinions about whether or not these protests are good or bad, but plenty of people across the world would love to be able to speak their mind at their job without it putting their career in jeopardy.

The Bay Area is really just the perfect place for this because it is a highly liberal area (in some aspects, anyways) and people feel extremely safe. So you get these sorts of liberal protests. I'm sure it would play out the same anywhere if this set of conditions were true regardless of political leanings.

I work in the midwest. Tech jobs aren't rare in the larger urban areas, no differently than California. Drive south towards Fort Hunter-Liggett or toward Napa Valley and even in California tech jobs become rare. https://en.wikipedia.org/wiki/Plano,_Texas#Headquarters_of_m...

I do agree though that the ability to protest and boycott is an entitlement. How many security, facilities, janitorial, or vendor staff have that freedom at the Google campus? In that consideration it does not sound like such a campus achieves equal opportunity employment in theory, but perhaps might by contracting other those positions to service providers. If that is true and those service providers were to boycott a Google business priority it should be no surprise when Google would refuse to continue services from such vendors. Again, this is a horrible way to do business.

> I don't have strong opinions about whether or not these protests are good or bad, but plenty of people across the world would love to be able to speak their mind at their job without it putting their career in jeopardy.

I vote with my time. If I become disgruntled with an employer I find another. Tech jobs, even in the midwest, are always in crazy demand for talent. I certainly don't put my employer's business in jeopardy to enjoy externalizing a personal opinion.

> I'm sure it would play out the same anywhere if this set of conditions were true regardless of political leanings.

History does not reflect this. I am old enough to remember the bay area was just like this during the 80s since I lived in Santa Clara. More lucrative markets were not like that without regard for job security, such as the financial district of New York. Other geographies are not like that. Most people are not willing to risk their employment relationship for the sake of freedom of exercising politics and actually look upon as a negative and potentially damaging behavior. To make the assumption that other areas are likely reflective of one's own is a form of ethnocentricity - https://en.wikipedia.org/wiki/Ethnocentrism

>I do agree though that the ability to protest and boycott is an entitlement. How many security, facilities, janitorial, or vendor staff have that freedom at the Google campus? In that consideration it does not sound like such a campus achieves equal opportunity employment in theory, but perhaps might by contracting other those positions to service providers.

I believe you are correct about all of that.

>If that is true and those service providers were to boycott a Google business priority it should be no surprise when Google would refuse to continue services from such vendors. Again, this is a horrible way to do business.

What's good business is very much opinion, many people would work for less money to keep their dignity. It's not exactly a Silicon Valley secret that Google is not the highest paying employer, after all.

>To make the assumption that other areas are likely reflective of one's own is a form of ethnocentricity - https://en.wikipedia.org/wiki/Ethnocentrism

When I said "anywhere" I meant "in the United States." Personally, I believe culture in the U.S. is more similar than it is different. This point could be debated but there's not much of a point since I doubt there's an agreement to come to with that.

I also was not intentionally referring to all types of jobs, just software developers. It obviously doesn't make as much sense with the power dynamics of jobs that are seen as unimportant. Programmers also take months to get ramped up in a company, and carry a lot of trade secrets and knowledge that they will take with them when they leave, making it very different from a line cook for example. Therefore, you probably don't want to fire a programmer if you don't have to, and you especially don't want a bunch of programmers with important domain knowledge to all leave at the same time.

So quite a bit less broad than "anywhere."

Still, I do not have as much life experience as you, and I do not know why Silicon Valley was different in the 80s. If I had to guess, I'd guess it's because there was no Internet, and the job market was insanely different. Today most software companies have standardized on things, but in the 80s there were plenty of different microarchitectures and programming languages you might need to know on top of other mandatory skills that are not so common anymore. Nowadays for most software developers you don't need to care about microarchitectures and everyone uses languages that look like C or JavaScript. The big tech companies were definitely a lot smaller. Employees sometimes even had more direct say than they do now. Reading old stories about Apple on Folklore.org or Quora paint a picture of an employer that employees respected, but disobeyed frequently. But if they wanted to protest, would anyone hear them? Apple wasn't the massive company it is today in the 80s. Without the Internet, their message might've gone unheard entirely. And, without the Internet, a lot of employees might not have even known about a DoD deal or the implications, positive or negative, that it might have.

Basically, comparing the tech industry of the 80s to the tech industry of today feels a bit strange. Today the tech industry is one of the world's largest. In the 80s, it was very impressive if you shipped 1000 computers.

> What's good business is very much opinion, many people would work for less money to keep their dignity.

I write JavaScript for living and I have mostly worked large monolithic Java shops, so I know this is a lie, at least in software. Most software developers are simply happy to press buttons against a framework while continuously achieving a paycheck. The thought continuous refactoring or doing it free, as in open source contributions on your personal time, is still a complete mystery to many corporate developers.

Dignity is certainly not immediately present in much of the code I have seen in my career. If dignity were the principle aspiration people would not be selective in its application. I suspect the motivation is something more immediately socially conforming. Contrarily, dignity is often expressed by rejection to conforming social pressures.

> But if they wanted to protest, would anyone hear them?

The bay area has never had a problem with this and in the past didn't need social media to prove it. More traditional media channels were sufficient. The area is substantially more sensitive and less diverse now though.

>> What's good business is very much opinion, many people would work for less money to keep their dignity

> [...] I know this is a lie, at least in software. Most software developers [...]

Many != most, and even if it did, different interpretations != a lie.

Unsure if you meant the implication, but north Texas is most definitely not the midwest.
Why not? Similar location. Similar geography. Similar ecology. Similar climate. Similar economy (cost of living and infrastructure). Similar weather, though north Texas is the southern extreme of it.
> Google employees don't get to choose their customers.

This theory is completely at odds with what actually happened. The reality is that workers at Google wield enormous power when they act together, and get to make the company drop a customer.

In that case why would that impacted customer want to do business there again considering that risk?
Because they should analyze why workers objected to the other project and understand the motivations rather than assume a blanket mistrust
And when the Pentagon analysis concludes that Google doesn't stay committed to military projects, what would they do?
Then they shouldn't use Google for military projects, which Google has already said they won't pursue (AI for weapons). But DoD =/= military, and if you are equating them here then you ignored what I wrote and are making a blanket statement rather than analyzing the specific case.
I'm confused. The article states that "JEDI is broader and bigger, seeking a vendor to provide cloud services for all branches of the military" but you seem to be saying it's not for the military.
Sorry for being imprecise, when I refer to military I mean specifically warfare. As the Google letter notes:

> > We believe that Google should not be in the business of war. Therefore we ask that Project Maven be cancelled, and that Google draft, publicize and enforce a clear policy stating that neither Google nor its contractors will ever build warfare technology.

So the issue is specifically warfare technology, not DoD/military bureaucracy in general.

not DoD/military bureaucracy in general.

The bureaucracy exists solely to get warfighters downrange and on-target. I mean, what do you think they do all day, and why?

I would say collaborating on a Google Doc about strategy is pretty far removed from powering the drone technology used to kill people even if that's what the doc is about.
Sure, if you follow all the links of causality, economic activity always produces federal taxes, which exist solely to fund the bureaucracy, which exists solely to get warfighters downrange and on-target. Therefore, nobody should ever spend or make money, lest you be supporting an immoral regime dedicated to killing people.

There are a few zealots that actually think like this, but most people realize that each stage in that chain is pretty lossy. There's a good portion of your wages that are not taxed, and there's a good portion of your taxes that are not directed to the DoD. Similarly, there's a good portion of the DoD bureaucracy that does not exist to get warfighters downrange and on-target - think of all the folks involved in procurement projects that are basically employment projects for defense contractors, or DARPA and the basic research it does, or the combat engineers building bridges in warzones, or the folks responsible for administering the VA and military pensions. You can be morally opposed to writing software that helps target drones while still believing that former vets should have access to health care.

there's a good portion of the DoD bureaucracy that does not exist to get warfighters downrange and on-target

No, actually there aren't. Whether or not they are successful or efficient is irrelevant. The money funding both the super efficient successful targeting system and the failed warfighter support system are fungible and come from the same "color" or pot of money.

It's not zealotry, it's understanding the purpose and mission of the organization.

Also the VA isn't part of the DoD. It's a Cabinet level agency.

Frankly it's getting annoying listening to people in tech talk about Defense matters as though they are authoritative.

[1]http://acqnotes.com/acqnote/acquisitions/color-of-money

50-52% of US federal budget = defense budget, so...
DoD is quite literally = to Military in the US. Not sure what you mean here.

The "bureaucracy" of the Pentagon directly supports warfighting across the entire spectrum. Efficiencies in government credit card processing for an Army SSGT going on temporary duty to train on employing the MRAP, enable effective execution of wartime tasks. This is a well understood and ingrained sentiment.

The disruption to business is sufficient for blanket mistrust. This is referred to as reputational harm. The ethical and legal considerations are fully investigated by both parties before the contract is inked. In business parlance this is called due diligence.

Speaking from a contractual perspective what does a customer care about the service provider's employee motivations? How would the customer reach into the confidential internals of the service provider's office to qualify or answer that question?

The motivations are public: https://static01.nyt.com/files/2018/technology/googleletter....

> We believe that Google should not be in the business of war. Therefore we ask that Project Maven be cancelled, and that Google draft, publicize and enforce a clear policy stating that neither Google nor its contractors will ever build warfare technology.

In that one instance the motivation is known, but it is a horrible and dangerous blanket business requirement that cannot be enforced. There is no way that ends well for a business relationship. It is up to the employer to discern and address the motivations of its employees. When this responsibility falls on the customer the business ceases to have customers. It doesn't make sense for a customer to pay for a service and be charged with the management of their service provider's managerial concerns, because if that were the case the service provider would become an internal business unit.

What is interesting is that the letter says this: This plan will irreparably damage Google’s brand and its ability to compete for talent. The letter is completely addressing the concern for reputational harm, but the company did not do so prior to the establishment of the relationship as a matter of its due diligence considerations. If this is cause for the DOD to pass on Google's bid for a larger project the harm is absolutely manifest. This is analogous to the Streisand Effect where the harm was neither well known nor established but achieves both due to the complaint - https://en.wikipedia.org/wiki/Streisand_effect

Also, the DOD is the US military and those terms can be safely and definitively interchanged. I can affirm this as a military officer. https://www.defense.gov/About/DoD-101/

>Not matter how well intended those employees' motivations it likely cost their employer a substantial piece of this enormous profit pie.

Who makes everything for Google? Who makes everything for the shareholders?

The workers, the employees. If they don't want to work on something that is against their morals and ethics as a whole, they get to decide that.

Do you work at Google or is that an assumption?
It's true in any circumstance. If you get rid of the employees -- the workers -- who's left to make anything, clean anything, provide food, etc?

Capitalists somehow forget that the workers even exist and are a necessary part of businsses, and seemingly think that all work that gets done is by pure magic. It's an immoral and unethical to not care what the workers say in quest of your personal profit.

That in a world in which Azure, AWS and GCP exist there also exist people who would choose Oracle or IBM clouds literally boggles the mind. Who on Earth are these people?

Aliyun is just as bad of course but at least it’s users have the ultimate excuse

Are kickbacks a thing? It wouldn't surprise me Oracle and IBM nurtured that or similar behaviour, massaging client's decision makers like that.
Yeah... from an engineer’s POV or a publicly-visible commercial POV it makes no sense. Something fishy is going on for Oracle to even get a seat at this table with their sorry excuse for a cloud.
15 years ago I worked with Oracle and guys from Oracle on a marketing side (they contracted me to help out with their marketing side), I definitely got that vibe. Years later, there was this talk: https://www.youtube.com/watch?v=-zRN7XLCRhc And it kind of all makes sense.
It's completely absurd that anyone would choose IBM or Oracle, but having heard first hand the sales pitch that IBM cloud give you I understand it a bit given that a lot of people with decision-making power don't have enough tech expertise to detect nonsense and FUD when they hear it.

In my case, the guy spent half the pitch flat out lying to me about AWS and how Amazon charges etc. (sample lie "They basically don't let you transfer any data out once it's in the cloud"). I let him go for quite a while before I revealed that I actually know what I'm talking about and calling out his bs ("Gee that's pretty weird because I host quite a few servers in AWS and I'm able to get data from them just fine. I wonder how that works?")

> They basically don't let you transfer any data out once it's in the cloud"

And how much does transferring 1PB of your data from AWS to say GCP cost, with those outrageous egress costs?

That's how they stop you.

(comment deleted)
It's a name recognition thing. I've seen cases where people in decision-making positions are like "IBM? They're big and have been around for a while - they won't fuck this up."

Spoiler: they did.

Used to work at Oracle. Do not do business with them under any circumstance.

As far as corporate greed, inefficiency, terrible service and support goes, Oracle is the worst of all corporations.

A simple Google search for how Oracle extorted $150M USD from the State of Oregon and flat out lied to them about their implementation services should provide all you need to know about how Oracle treats its customers.

My attempts to rectify these problems got me fired from Oracle. It is political and it is a cash grab.

Do not go near Oracle.

> A simple Google search for how Oracle extorted $150M USD from the State of Oregon and flat out lied to them about their implementation services should provide all you need to know about how Oracle treats its customers.

I did that simple Google search. I read that Oregon demanded to own the project coordinator role, with ultimate authority, and only hired Oracle for labor and machines, against Oracle's advice, and then failed to actually hire an individual to fill that role. Oracle's hands aren't clean here, but don't whitewash Oregon's obvious screw ups.

Oracle claimed to be providing a team of software engineers who were coding Oregon's state exchange. Oracle was much more than a vendor, it was a service provider.

Oracle executives flew out to Oregon 3 times, and on each occassion, made presentations that said the exchange was "72% done" and that if they would just wire them 53 million USD, everything would be done on time.

These were clear and deliberate lies, and as taxpayers, one should feel personally robbed by Oracle, regardless of the State of Oregon's overseeing the project.

IBM is another monster delivering jack squat for big dollars. Canada spent over $1B CAD on a deployment with them for a new payroll system and ended up with diddly squat for that much taxpayer dough.

https://www.pymnts.com/news/b2b-payments/2018/ibm-canada-pay...

I couldn't agree with you more. These companies have become such a terrible value for consumers that they have to abuse the public trust and effectively win business through lobbying and nepotism.

And lots of golf.

I dont know. From what I know from inside automotive, the IT departments there are buying new compute clusters about every second year.

After they make their requirements public usually a few companies show interest. The companies swear by their techs and claim it tested for requirements, yet almost always this new company have to back up after they prove themselves unable to keep up to their promises.

The IBM was the one delivering past several years of compute hardware, while other companies were just not able to do the same.

Claiming whole corp is rubbish because you know of one case where it didnt work out seems not right here.

Remember the old adage, nobody ever got fired for buying <x> or in this case IBM.
It's enough to see how they're killing Java's ecosystem, moving to the 6-months release cycle.

There's still a huge amount of companies / public businesses running software based on Java 5 or 6 out there, just saying..

That's partly on companies too. The fact that software marches forward has been clearly understood for decades and keeping things updated should be part of IT/Dev responsibilities, not something put off forever.
Given Oracle's relative small Capex investment in cloud, one wonder if this is the only contract they want.
(comment deleted)
Should a single cloud supplier be considered a single point of failure in the event of war?
I think it's possibly a single, point of subversion by foreign intelligence.
Similar to say.... a president?
Abolish the presidency!
(comment deleted)
Unless the war is between the cloud supplier and the gov't, I don't see this being an issue. The standard redundant physical location policies already in place should protect against a facility suddenly becoming catastrophically offline.
Physical redundancy doesn't do much if your infrastructure is controlled by software (deployment, monitoring, scaling, pipelines for updating dependent services) that isn't well controlled.

GCE and AWS have both suffered widespread connectivity issues due to deployments. Adversaries don't need to blow up data centers, they just need to modify the code used to provision, monitor, and control them.

So what if tech companies can bid? It’s $10B, they likely already know who they’re going to give it to. Most of this stuff is wired anyways.
There's usually a fulfillment requirement - basically proving the company can handle contracts of this size, often from prior contracts.
Right. Of course there is. That’s not my point. My point is that, before accepting any applications, the administration likely already knows who they want to give this contract to. That entity will be capable of doing the work, of course, but there’s too much politics in $10B for it to be given solely on merits.
The terms are also important, and assigning the contract without any competition ensures they will be incredibly one sided.
Yep they use it to drive up the price of the contract.

Reminds me of when I was a pledge a decade ago. The fraternity with one of the best houses was kicked off campus. So the owner, seeking new occupants, put the house up for bidding. There were a lot of implications to having your letters on that house— anyone who had that house would automatically be guaranteed an excellent fraternity experience. Everyone who was in the know already knew the seller was going to give it to one specific fraternity that wanted it. So most fraternities who were eyeballing the house didn’t even bother. Still, my fraternity engaged, and in the end it was the remaining bidder along with the aforementioned. Ultimately, even though we gave the best bid (we could fill up all the rooms and at a better price), it still ended up going to the other fraternity, because politics. However the biggest winner was the seller, because they not only gave it to who they wanted to give it to from the start, but they also got to drive up the price for it.

Even though that was my first real exposure to this kind of stuff, I’ve seen over the years that this is par for the course. Actually just last month my company, partnered with two universities, was involved in what turned out to be a wired $10 million grant. Truly open bids are rare for contracts with major implications. For any big contract, it’s good practice for interested parties to use their connections to try to find out if it is wired or truly open before spinning their wheels and dumping their resources into it.

As long as Amazon doesn't get it, I'll be pleased.
Doesn't Bezos already have a seat at the Pentagon? I thought he was locked in. Maybe that was the CIA I'm thinking of.
This kind of contact makes no sense!! Even an iPhone has alternate vendors for every external part used, how can the Pentagon award it to a single vendor?

They need to pick 3, in my opinion, and make the 3 constantly compete against one another for higher shares in profits and revenues - you know - an open market