2. Why would the ex-CIA guy hurt his current employer to help his former? Don't people do it the other way around?
3. Good luck finding a single Fortune 500 company that doesn't employ someone that used to be in the Intelligence Community.
4. Telnet options are pretty arcane. It seems like an easy bug to write. The modus operandi of U.S. backdoor attempts in the past matches a different model. U.S. backdoor attempts fit two models: backdoor dependent on secret key (Dual_EC RNG), and backdoor dependent on physical sabotage (interdiction). This is stuff available in the Snowden docs and related news reports.
It's a common HN meme that Cisco and Microsoft help the U.S. government spy, but there isn't credible evidence supporting it. They actively resist government espionage attempts.
Yahoo, RSA, et. al. deserve the negative attention, not companies that fight the good fight.
This is actually a completely fair point...
Look at all the ex SS at FB.
Microsoft was the only company featured in slides about teamwork and collaboration.
https://www.theguardian.com/world/2013/jul/11/microsoft-nsa-...
When you get into the big wide world of making billions of dollars, you'll find that people are more than willing to throw you a bone long after you're gone to make sure you're still invested.
I've seen this first hand (unfortunately/fortunately, not the recipient of said bone).
I don't think quitting your CIA job lets you off the hook for essentially giving away something they consider to be a secret. Or at least I'm sure the CIA would not look at you kindly for doing so.
Those links were the result of a few cursory Google searches. It wasn't intended as a comprehensive representation of Cisco's CIA ties, nor to imply that anyone mentioned was directly involved in backdoors. Rather, the point was that when a security-oriented organization employs ex-spooks, it increases the likelihood of spooky things happening. At the very minimum it fosters distrust.
>2. Why would the ex-CIA guy hurt his current employer to help his former? Don't people do it the other way around?
While I didn't mean to imply any specific individual, I'll address the underlying rationale: When someone's former employer is a powerful government spy agency built on secrets, the loyalty dynamic is a bit different than that of normal private sector job hopping.
Moreover, I'm sure the CIA isn't stupid and tends to maintain relationships—intentionally or not—with former employees that cross over to private sector work, especially high-level people since they're already trusted, reliable, and may be taking positions that later prove to be beneficial to CIA interests in the future.
>3. Good luck finding a single Fortune 500 company that doesn't employ someone that used to be in the Intelligence Community.
This is a fair point. Given the nature of Cisco's products, and combined with recent leaks detailing the pervasive nature of US offensive cyber efforts, it should raise more suspicion than normal when former IC people go to work for a company dealing with network hardware.
That said, I'm not sure how you'd quantify the number of former IC people Cisco hires versus a normal Fortune 500 company. For all I know it could even be lower than average.
>... and backdoor dependent on physical sabotage (interdiction) ...
This approach is far superior to baking vulnerabilities into entire product ranges (or even standards). It doesn't compromise the private sector in the process, beyond perhaps supply chain integrity. The flipside of course is that physical implants don't really scale well, but at least the blowback is relatively minimal when things go wrong. Not to mention it's just badass in a James Bond kind of way.
>It's a common HN meme that Cisco and Microsoft help the U.S. government spy, but there isn't credible evidence supporting it. They actively resist government espionage attempts.
While that's true, it's also possible a company can publicly say or do one thing—genuine or not—and actively do another because of either connections, or that they're being compelled against their will by NSLs. It's also possible to have factions or even individual employees within a company that are actively subverting security with the addition of backdoors, unbeknownst to management.
> it should raise more suspicion than normal when former IC people go to work for a company dealing with network hardware.
Fair enough about the perception of former IC employees entering the private sector. For the most part, "former IC employee" only impacts an organization once something else newsworthy occurs. Cisco considered the risk to be low that the CIA would pay a former official to order Cisco employees to insert a backdoor without a single one reporting this outside their chain. Personally, I think it benefits companies to hire employees with a diversity of backgrounds.
> [backdoor] approach is far superior ... [they] just don't scale well ... it's just badass.
The lack of scalability is a feature if you ask me. It should be expensive and difficult enough to limit espionage to those target that matter. And it's more palatable to the people who these agencies ultimately report to.
why? how is someone who has been vetted by the IC as a trustworthy keeper of their secrets become less trustworthy?
> CIA isn't stupid and tends to maintain relationships with former employees that cross over to private sector work
what kind of "relationships" are you talking about here?
Because their loyalties may lie with their former employer. National security matters tend to take priority over private sector loyalty for many people, and I'd imagine especially so for former IC types since they tend to be extremely patriotic. I mean that as a compliment, not an insult.
>what kind of "relationships" are you talking about here?
Personal relationships. Say Joe retires for private sector work. He's good friends with Bill, who's a case officer. Later when some national security matter comes down the pipe requiring access at Company Y, Bill knows Joe who works there. Seeing that Bill and Joe have been through some shit together, that relationship is probably going to supersede loyalty to any private sector employer when there are underlying national security motivations involved.
It's not tinfoil hat conspiracy allegations, just basic human dynamics.
> Because their loyalties may lie with their former employer.
By that logic, Apple shouldn't hire any employees from Google, because they might reveal iOS secrets.
> Personal relationships... that relationship is probably going to supersede loyalty to any private sector employer
Sure, anything could happen. In your example, Joe gets nothing except enormous risk of public humiliation and expensive litigation in exchange for the favor to Bill. Any employee that is involved in this would almost certainly speak to someone. This isn't like a National Security Letter; there's no force of law to compel silence. It would be a foolish move for Joe.
Then Bill gets a call on his gray phone from his boss. "Why is the CIA in the newspapers for attempted backdooring of Cisco products?" We don't hear about this stuff because it doesn't happen like this.
People that are so loyal to the CIA don't leave for a pure private sector company. They go to a contractor so they can stay in the ecosystem. People going to Cisco are (1) physically moving away from Northern Virginia, (2) losing access to secrets.
Besides, it doesn't sound very patriotic to insert backdoor that's not even protected by a secret (e.g. key escrowed Clipper Chip, P and Q in Dual_EC) that goes into hardware purchased by more Americans than anyone else in the world.
Apple and Google are on the same playing field though. They're fundamentally the same type of entity. An intelligence agency vs a pure private-sector company is not.
>Any employee that is involved in this would almost certainly speak to someone. This isn't like a National Security Letter; there's no force of law to compel silence. It would be a foolish move for Joe.
While certainly risky, I'm not saying Joe wouldn't be legally protected. His favor very well could be under the auspices of an official program. Heck, IANAL but it possibly could even be in the form of an NSL. As far as I'm aware they're the legal equivalents of blank slates and don't necessarily require informing the upper echelon of a company of their issuance, but I could be wrong on that. I know that it is at least customary to do so, however.
Either way, let's say the government wants to issue a NSL requesting a very specific, perhaps even temporary backdoor to a Bay Area company. The normal route would likely start a veritable war with that company's legal department, and runs a non-insignificant chance of being leaked by those who know about it.
Contrast that scenario to issuing the NSL directly to Joe, who they know in advance is solid. If Joe happens to be an engineer, mission accomplished. If he isn't technical, then maybe he just happens to bring a USB stick to the office one day and plugs it in.
If that sounds far-fetched, keep in mind that the NSA compromised Google's internal network during the course of their operations, so network exploitation against U.S. companies isn't even off the table.
>People that are so loyal to the CIA don't leave for a pure private sector company. They go to a contractor so they can stay in the ecosystem. People going to Cisco are (1) physically moving away from Northern Virginia, (2) losing access to secrets.
All good points, but it doesn't mean they're cutting ties and shunning their former life either.
>Again, I'm glad you're engaging. I'm not down voting you.
Thanks. :)
Good point; my analogy fails.
> NSA compromised Google's internal network
News stories that used that phrasing were being inaccurate; the collection was of plaintext traffic between international Google datacenters. The Intercept explains it pretty well.
> As far as I'm aware they're the legal equivalents of blank slates and don't necessarily require informing the upper echelon of a company of their issuance, but I could be wrong on that. I know that it is at least customary to do so, however.
The point I'm making is that it's rational for people not well-versed in the minutiae of foreign intelligence regulations to think that intelligence agencies can and do intrude however they please on private companies. In reality, its with great caveats and with a (perhaps insufficient) legal process that introduces friction. It is far more overt and less cloak-and-dagger than would ordinarily be desired. The most clandestine way possible to spy would be with the badass methods we talked about earlier.
> Either way, let's say the government wants to issue a NSL requesting a very specific, perhaps even temporary backdoor to a Bay Area company. The normal route would likely start a veritable war with that company's legal department, and runs a non-insignificant chance of being leaked by those who know about it.
NSLs aren't a magic spell that gives the government whatever it wants. They go through a company's legal department and are subject to limitations. Several companies have successfully defeated the gag order portions of their NSLs. Having an inside man wouldn't be particularly advantageous since the company lawyers must get involved (the U.S. hasn't yet prohibited a person's legal representative from viewing writs).
At the end of the day, the government has a monopoly on violence and can sent the Army to invade Cisco or anyone else it doesn't like. But due to the conscientious individuals that make up the government, and the public relations and legal risks that are involved, this is unlikely to happen. What I'm describing is a series of norms that are unlikely to be broken in the course of spying business. There are far cheaper, easier, and less risky ways to accomplish espionage than a covert, mass-distributed, plaintext backdoor.
I do wish that the credibility of U.S. companies was taken into account more often when espionage decisions are made. The terrible optics behind PRISM made it look like companies were volunteering information, when the government was coercing them through the NSL program.
A US intelligence agency targeting a portion of a US company's infrastructure that just happens to be international still constitutes a breach of trust. Whether that means they'd go as far as compromising an office network located in the United States is another matter, but I seem to recall a talk given by a security chief at Google discussing their use of custom RISC-V silicon for security due to their threat model including nation-state actors, with explicit mention of Western intelligence agencies. I wish I could find the link.
>The point I'm making is that it's rational for people not well-versed in the minutiae of foreign intelligence regulations to think that intelligence agencies can and do intrude however they please on private companies.
The mental gymnastics[0] that the NSA already uses to legally justify domestic collection on US citizens doesn't really inspire confidence that the IC wouldn't stoop to similar antics when it came to private companies.
>Having an inside man wouldn't be particularly advantageous since the company lawyers must get involved ...
Must they? My point was that legally compelling an insider who's already on your side might serve as an interesting loophole. Especially if they're not required to inform counsel and have no intention of doing so. Like I said though, IANAL.
>... (the U.S. hasn't yet prohibited a person's legal representative from viewing writs).
It's worth noting that in the early days that wasn't so clear:
"CONAN: And they are roughly equivalent to subpoenas.
Mr. LICHTBLAU: Yes and no. There are differences - one of the key differences is that for a long time the recipient was not even allowed to tell you when they received such a letter.
CONAN: Even their own lawyer.
Mr. LICHTBLAU: There was debate about whether or not you could even get a lawyer." [1]
>I do wish that the credibility of U.S. companies was taken into account more often when espionage decisions are made. The terrible optics behind PRISM made it look like companies were volunteering information, when the government was coercing them through the NSL program.
Agreed.
[0] https://news.ycombinator.com/item?id=10605489
[1] http://www.npr.org/templates/transcript/transcript.php?story...
Eric Grosse at 5th RISC-V Workshop, 2016:
https://www.youtube.com/watch?v=0knR6vXba7g
Slides: https://riscv.org/wp-content/uploads/2016/12/Tue1330-RISC-V-...
once any single person enters the IC, they have lost trust of anyone outside of the IC... simple - and answers your other question as well...
https://tools.ietf.org/html/rfc748
NWG/RFC# 748, M. Crispin, 1 April 1978
Telnet Randomly-Lose Option
1. Command name and code.
RANDOMLY-LOSE 256
2. Command meanings.
IAC WILL RANDOMLY-LOSE
The sender of this command REQUESTS permission to, or confirms that it will, randomly lose.
IAC WON'T RANDOMLY-LOSE
The sender of this command REFUSES to randomly lose.
IAC DO RANDOMLY-LOSE
The sender of this command REQUESTS that the receiver, or grants the receiver permission to, randomly lose.
IAC DON'T RANDOMLY-LOSE
The command sender DEMANDS that the receiver not randomly lose.
3. Default.
WON'T RANDOMLY-LOSE
DON'T RANDOMLY-LOSE
i.e., random lossage will not happen.
4. Motivation for the option.
Several hosts appear to provide random lossage, such as system crashes, lost data, incorrectly functioning programs, etc., as part of their services. These services are often undocumented and are in general quite confusing to the novice user. A general means is needed to allow the user to disable these features.
5. Description of the option.
The normal mode does not allow random lossage; therefore the system is not allowed to crash, mung user files, etc. If the server wants to provide random lossage, it must first ask for permission from the user by sending IAC WILL RANDOMLY-LOSE.
If the user wants to permit the server to randomly lose, it replys with IAC DO RANDOMLY-LOSE. Otherwise it sends IAC DONT RANDOMLY-LOSE, and the server is forbidden from randomly losing.
Alternatively, the user could request the server to randomly lose, by sending IAC DO RANDOMLY-LOSE, and the server will either reply with IAC WILL RANDOMLY-LOSE, meaning that it will then proceed to do some random lossage (garbaging disk files is recommended for an initial implementation). Or, it could send IAC WONT RANDOMLY-LOSE, meaning that it insists upon being reliable.
Since this is implemented as a TELNET option, it is expected that servers which do not implement this option will not randomly lose; ie, they will provide 100% reliable uptime.
Once you're a made man in an intelligence service, you're in for life, regardless of whether or not your name appears on the official payroll.
I doubt ThreatGrid is the only IQT company Cisco has bought. I am as certain they're not a CIA plant inside Cisco as I am of anything else in my perceived reality; I know several of their key people, one of whom has a framed Che poster. They're not the type. :)
You're going to drive yourself seriously crazy if you try to trace every IQT investment that's been acquired by any company and then game them out as "compromised by the CIA".
Of course they aren't. All I'm saying is if you have a company that has a routine acquisition deal flow from the CIA's venture capital arm, that there's probably healthy working relationships there.
I think you just can't read anything into In-Q-Tel involvement with a company.
Perhaps the plan with In-Q-Tel was partly to achieve pervasive private sector influence, and it was somewhat successful.
While I appreciate the conceited scare quotes there, I'm not saying that working relationships directly translate to backdoors, just that it translates to influence, and that can indirectly lead to bad things over time.
>I think you just can't read anything into In-Q-Tel involvement with a company.
As a thought experiment, say Open Whisper Systems took money from In-Q-Tel, or even acqui-hired a team from one of their portfolio companies. Right or wrong, people would very much read into it, and probably lose their minds in short order. Signal's user base would plummet overnight.
I'm not sticking up for In-Q-Tel. I wouldn't take money from IQT. But it is, for the most part, just what it says it is: an investment firm that funds things that the CIA believes will be helpful to its mission, which includes virtually all network security, database, networking, and RF technology.
Wasn't my intention for that comment to be deceptive in its implication. After seeing how some of the replies were going, I did try to amend it in a more balanced fashion—then ended up having to move that further down the thread due to the edit window expiring.
Funny enough the original comment's karma continued to skyrocket while the more measured words teetered on the brink of downvote for some time. It was surprising, since usually HN is pretty good at punishing anything that even has the slightest perception of hyperbole.
That said, I think we can agree that the 2-hour edit window represents a grave threat to civil discourse, freedom of expression, also freedom itself, probably national security, and nothing less than the future of Hacker News. It must be extended. :)
Perhaps a bit OT, but you're dead wrong about Che. He was a mass murderer, and I'd treat someone with a framed picture of him with the same suspicion I'd reserve for someone with a Stalin or Hitler portrait.
"Cisco vice president of services Mike Quinn, a former CIA operations officer, ..." [0]
"... Cisco's recent acquisition of In-Q-Tel-backed security company ThreatGRID ..." [1]
"After retiring with 30+ years of service from the Agency, I spent several years as adviser to Cisco System’s Chief Security Officer, and I found Cisco was doing great work; they’re one of the backbones of the internet." [2]
Imagine that.
[0] http://www.networkworld.com/article/2223473/data-center/cisc...
[1] http://www.networkworld.com/article/2358453/security/cisco-p...
[2] https://www.forbes.com/sites/realspin/2012/10/12/what-do-for...