The US Senate just approved Signal for staff use. Signal is a secure messaging app with no backdoor, and no large corporate owner who can be pressured to install a backdoor.
Susan Landau comments.
Maybe I’m being optimistic, but I think we just won the Crypto War. A very important part of the US government is prioritizing security over surveillance.
Just An Australian • May 17, 2017 3:09 PM
Don’t count your chickens before they hatch – Australian politicians have been using Signal and similar apps for a while already, to avoid surveillance, even while they vote for increase surveillance over the citizens that have to vote for them
Winter • May 17, 2017 3:11 PM
“Back to being pessimistic, Any Half-Decent Hacker Could Break Into Mar-a-Lago”
Why is that pessimism?
As Trump shares all the valuable secrets with the wrong people, why should the rest not be informed too?
JG4 • May 17, 2017 3:32 PM
who actually has endpoint security? maybe someone using a data diode, but not many others
JW • May 17, 2017 3:53 PM
Also, what does surveillance mean in this case? It could easily mean “oversight by citizens”. As I understand it, Senate staffers only have to archive and make available things that are ‘historically significant’. Who makes that call? And how can we tell who is making it and why?
Pete • May 17, 2017 3:54 PM
Congress has a habit of making different rules for themselves vs everyone else.
Rhys • May 17, 2017 4:04 PM
The likening to the anecdote about a knife being a tool of great good in the hands of surgeon; that same tool in the hands of child is but regret… comes to my mind.
Encryption without understanding classification just invites more problems than it solves.
As to an alleged crypto war- privacy with encryption is likely a Pyrrhic victory at best. Or does it just really mean Phil Zimmerman can finally come home?
IMHO, surveillance is balancing act. Not mutually exclusive with privacy. (Audits are a surveillance, particularly outside or 3rd party audits which serve a valuable function.) With the Patriot Act that public has been sanguine about, surveillance is still at the Government’s discretion.
Between FOIA and the Sunshine Act, do you suspect that elected or appointed staff, let alone a significant plurality of users, if tested would know when to use what? (Although defining “historical significance” by the WWRMWD, what would Rose Mary Woods do rule, is likely the prevailing outcome.)
What is your understanding of this body’s application of the principle of accretion?
Or will elected and appointed staff just defer and use encrypted form regardless of the content?
Since we already have recent examples of elected and appointed officials’ spills (leaks, errs, other equivocations)- how many would confuse that ‘Signal’, without respect to Snowden’s endorsement, is/isn’t acceptable for sensitive/classified data?
All parked on devices with discretionary access controls.
Bob • May 17, 2017 4:20 PM
Is/Are Congress & Congressional employees still required to archive all communications? I thought public work required public records? Perhaps I don’t understand how Signal works.
Jack • May 17, 2017 4:29 PM
@Bob, read the Engadget link that is in Bruce’s blog post. Apparently not is the answer to your question.
Who? • May 17, 2017 4:53 PM
What Senate? The same Senate NSA was spying? Or it was the Congress?
The very same Senate whose ID SmartCards use a photo of a chip instead of a real one?
I am sorry, U.S. Senators are not what I would call experts on cybersecurity. Their technical opinions have no value, only their political decisions have.
Who cares if Signal is secure or not? Is it one of the tools Snowden recommends? I guess it is, but it does not make Signal secure. It runs on the most vulnerable operating systems that exist right now. Android? What a joke. iOS? Don’t ask me. Windows? Linux? If the operating system is broken then the security of the app is broken. Am I wrong?
As Snowden documents shown a lot of times, NSA is usually not interested in breaking cryptography, it is computationally expensive and there are much easier ways to get to the data. Android, iOS, Windows, Linux… all these are easy targets. Operating systems developed by corporations that do not care about privacy nor security, or written by developers that do not care about security and whose chaotic (“bazaar”) development model makes very easy for the NSA sneaky add developers that carefully write backdoors.
Sorry, Bruce. We have not won the Crypto War. In fact, believing we have won it is the most dangerous movement we can do. We are not safe.
Who? • May 17, 2017 5:05 PM
@ Just An Australian
Don’t count your chickens before they hatch – Australian politicians have been using Signal and similar apps for a while already, to avoid surveillance, even while they vote for increase surveillance over the citizens that have to vote for them
It is sad when politicians of one of the five eyes need to protect themselves against their own surveillance network. Something is really broken in our world.
ab praeceptis • May 17, 2017 5:16 PM
It’s mainly politeness that keeps me away from commenting on Bruce Schneiers ” We have won the crypto war” statement.
I’d strongly suggest to stay cautious and to wait with the victory parade…
And, uhm, certain doubts (am I diplomatic, or what?!) re. signal are just one reason for me being less confident of the victory.
As long as there are people like mc cain’t (living proof that brain dead wannabe rottweilers still can bark) in the senate no technology can possibly keep the gov. institutions out of trouble.
In fact, it seems to me that right now a major part of senate and congress is by far more dangerous for the us of a than Russia or China.
Most importantly, though, one can’t cure the plague with a band aid, no matter how great that band aid may be. In other words: The abysmal situation we’re in re. safe IT systems is not to be cured by some golden sticker super-software. To cure that disease a major cultural change will be needed.
We’ll be secure when the general situation is good enough to not need some golden sticker stuff like signal. Until then the productions rate of the plague abyss called software development will produce trouble way faster and more than some well-meaning activists(TM) can produce golden stickers.
Sorry for spoiling the victory parade.
D.A. • May 17, 2017 5:25 PM
Individuals should not be allowed to use encrypted communications unless they have a demonstrable cause. There is no legitimate reason for anyone worth less than seven figures to use encrypted communication. Their communications are of no consequence, and encryption will only complicate the jobs of our esteemed officers of the law. Encryption is a munition and must be controlled. This is no different than denying access to firearms to select members of the population.
supersaurus • May 17, 2017 5:39 PM
@D.A.
“…There is no legitimate reason…”
what country are you from? or were you being satirical?
Clive Robinson • May 17, 2017 5:50 PM
@ JG4,
who actually has endpoint security?
Not anybody using signal on a mobile phone, that’s for sure.
EvilKiru • May 17, 2017 5:56 PM
@D.A.
Why 7 figures? So that those worth 7 figures or more can worry less about whistle blowers ratting them out?
P.S. The only people who consider encryption to be a munition are craven cowards and their toadies.
Congress Finally Understands Privacy • May 17, 2017 6:00 PM
Dark Money Paid Political Ad
Whats wrong with using WhatsApp? Help Mark Z’s angry legions to replace Trump with Numero Uno 2!
WhatsApp designed for government compliance:
https://theintercept.com/2016/06/22/battle-of-the-secure-messaging-apps-how-signal-beats-whatsapp/)
Full Circle: The Hand Bites Back
Congress recently gave the FBI broad legal authority to hack anyone simply for using a VPN or Signal encryption. So, like, can Mr Trump legally tap lawmakers ‘wires’ to see whose leaking?
We Have No Secrets
Have lawmakers been able to see THEIR highly detailed Big-Data dossiers? Are they too destined for hacking like OPM dossiers?
Do they still carry their personalized tracking/eavesdropping devices into meetings and the hallways?
Marked Increase of BS
Do addicts become highly irrational when denied the extension to their arm? Do they frequently contradict themselves and send employess running for the bushes?
Trip to Vegas
Do data-mining engineers program products to never put the gambling devices down?
Today & Tomorrow
All this is America today. If you thought Millennial’s were bad just wait for the Snowflakes…
milkshaken • May 17, 2017 6:02 PM
It is just this round of crypto war, within 5-10 years the national security state will come with new demands on making phones and laptops decryptable. Some things never change.
Wannacrypt+Shadow Brokers dumps discredited government demands for NOBUS back doors. Even Microsoft, who must have known about unpatched vulnerabities being exploited by USGovt agencies, and kept them unpached for business reasons and out of respect to USGov, will think twice about de facto backdooring Windows in the future.
Rex Rollman • May 17, 2017 6:13 PM
Encryption will end up like insider trading: Congress can do it but normal people can’t.
Bob Dylan's Happy Face • May 17, 2017 6:16 PM
I’m less impressed by this development than @Bruce is. As a former Senate staffer myself (who became interested in cybersecurity long after I left DC) the value of what Senate staffers talk about is probably less than what the popular imagination might expect it to be. Mostly, staffers are dealing with constituents (broadly defined, to include lobbyists of various stripes) about issue that are public or working with the Senators themselves, again about issues that are public. The “sensitive” stuff is really about political strategy in elections or strategy in terms of vote counting in the Senate itself. There is a reason why they are exempt from most public records laws and that is because they often are not involved in anything interesting, with interesting to be understood as national security drama. So in my view this development is akin to allowing teen girls to encrypt their messages–it is useful insofar as it sensitizes them to security related issues but no one besides themselves really cares about the make-up and the boyfriend drama.
From a security related viewpoint, I am also less impressed because as other posters have pointed out there is the huge problem of endpoint security. Which weakens this development considerably.
So sure this is a small and positive PR coup for the encryption side of the debate but I wouldn’t read as much into this development as @Bruce does.
Why is it so good that Senate can make secret deals more easily? Shouldn’t we be working towards more open lawmaking and govt, not more secret deals?
Anura • May 17, 2017 6:24 PM
It seems to me that they should be using something that has been backdoored so that just the NSA can decrypt it; the backdoor would probably not even be used unless they were actually doing something illegal in the first place. When I see them going to such great lengths to keep secrets from our government, I can only conclude that they must be conspiring to do something illegal; they may even be planning on overthrowing the government itself.
Bob Dylan's Happy Face • May 17, 2017 6:50 PM
@Roger writes, “Why is it so good that Senate can make secret deals more easily? Shouldn’t we be working towards more open lawmaking and govt, not more secret deals?”
Not sure if this reflects your thinking or your thinking about how others might perceive this development but either way it is wrong. This has nothing to do with the Senate making deals–for the most part Senators do not make deals via staff members. Deal making is done Senator to Senator, with the staff often finding out after the fact what deal is about to go down. Hell I would often not be told what was going to happen and learned it on the TV just like everyone else…and then be expected to deal with the pissing and the moaning of the various parties who felt they had been given the cold shoulder.
Again, I want to be careful here and not say that the data which is soon to be encrypted has no value. Sometimes it might be useful for a hacker to know metadata things like who is talking with whom. Relationship data can be powerful. Generally speaking, however, the data is simply the mundane data that is the equivalency of you talking with your wife about what is left in the fridge and what needs to be bought. No one cares, not even the Russians.
Dave Howe • May 17, 2017 7:05 PM
Sadly, no. The senate have always believed they should be exempt from the data gathering they feel is perfectly fine for everyone else. This is just re-enforcement of that.
Patriot COMSEC • May 17, 2017 7:30 PM
It is a interesting development. It means the someone in the Senate is aware of what is going on with collection. After all this time since Snowden spilled the beans, someone with a pulse and a living brain cell has decided to encrypt. Too bad they could not have done that for the Office of Personnel Management before the Chinese robbed the cookie jar, with glee and amused contempt.
No one likes being spied upon, and at this very moment one can bet that collection devices belonging to several countries are filling up with data from Senate devices. That task just got harder.
But don’t think for one moment that Signal actually works once it is used. The math is probably good, but the employment is surely prone to defeat. The idea of Signal is solid and convincing, its math is like the Titanic, but…
There are some good jokes hanging in the air. Remember how the CIA was spying on Diane Feinstein? Diane Feinstein, a lady who was very willing to allow spying on anyone and everyone in the U.S. –except her. Now it will be harder for the the deep state to spy on the Senate, and we should all applaud that. We are moving towards a bright future, as President Obama used to say, with a straight face, “Moving forward…” (which is, by definition, not backward).
Steve • May 17, 2017 8:11 PM
Yeah, hey, who would want public records to actually become public?
K.S. • May 17, 2017 8:18 PM
Information wars will not be won with cryptography. For cryptography to securely communicate between end points we need to have secure end points. This simply not true for all definitions of end points today. The only way to win crypto war is to poison the data – where it isn’t feasible to automatically tease out real data flow from the fake one. In that, encryption could only limit places to tap data flow to finite number.
Andrew • May 17, 2017 8:43 PM
Agree with what has been said, it’s more like we won the encryption battle but we kind of lost the war…
The Deep State is Evil • May 17, 2017 10:23 PM
Basically, these staffers are so concerned by the seep state, that they decided to use a technology not known to have been broken by them.
Make no mistake. It is not fear of Russian or Chinese eavesdropping that drove this decision, rather, fear of eavesdropping by the American deep state. Remember that Dianne Feinstein vigorously defended the NSA in the aftermath of the Snowden disclosures until she found out that her own committee had been spied by the deep state.
Encryption everywhere, Tor everywhere is the only line of defense we have against the who run the federal government.
Anon • May 17, 2017 10:33 PM
Waiting for the inevitable headline : “Signal outlawed except for official government use.”
kodozani • May 18, 2017 2:50 AM
What if NSA’s windsorGreen could break Signal Protocol so that they start using in order to give false sense of security.
Patriot COMSEC • May 18, 2017 4:15 AM
@ K.S.
That is right! Information wars will not be won by cryptography alone. Everything depends on the implementation. And you hit the nail on the head: “For cryptography to securely communicate between end points we need to have secure end points.”
The way to win the information wars is to air gap and encrypt and decrypt from an air-gapped device. That is what is meant by “secure endpoints”. The point here is that the end point then depends on the end-user. Why depend on a company unless you can read the mind of everyone in the company? You don’t know who is working for whom. Unter vier Augen: two people talking, four eyeballs, and you only have to trust the person you are talking to. That is the way it should be.
I agree with you in that the second prong of the attack is to poison the data. Make it fake.
People who want real security are going to have to admit to themselves that a device attached to the internet is unsecurable. The system is too complex to secure. The trail of electrons must be broken, upstream and downstream, and then best practices must be followed.
If you care about this fight, there are two things we need to encourage and do: (1) make it fake; make the data that everyone is fighting over worthless (2) air gap; teach people how to encrypt and move upstream and downstream from an air-gapped device
Also important are: hiding ciphertext, masking the amount of traffic, eliminating metacontent generally, and achieving anonymity
What Mr. Schneier is talking about here is that cryptography has been vindicated. If the Senate is doing it, then something is working. The Senate Intelligence Committee knows what works. But we very well know that cryptographic systems are subverted during their employment. Still, what Mr. Schneier has pointed out is good news, and it shows the value of his hard work.
Drone • May 18, 2017 4:44 AM
They’re taking it out for a test drive. If the like it, they’ll keep it for themselves and make sure you can’t have it.
Think of it like health-care, but for privacy. They get the best you can pay for – you get crap, and can’t afford it.
How to defeat all encryption
For Intel Platform:
– Leverage Intel AMT/SGX and AMD SP to meddle with userspace partition and escrow generated keys or mess up the randomness.
For ARM A series chip (found in almost every mobile platform i.e. smartphones)
– Leverage ARM TZ to meddle with userspace partition and escrow generated keys or mess up the randomness.Also note that ARM TZ architecture was the “grandfather” architecture that gave birth to Intel AMT/SGX, AMD SP, Samsung KNOX, Apple Secure Enclave and so on. Thus, the very nature of anything derived via TrustZone architecture is itself an inherent weakness despite it’s supposed goals of creating a “Secure Enclave” which follows a saying of “The road to Hell is paved with good intentions” and we can take the Secure Enclave design as a “supposed good intentions” but the result it procedures are hardware based DRM, OEM lockdowns, backdoors and so on.
It is not too far fetch to defeat all encryption by using such methods as mentioned above.
The NSA et. al. could continue to tap into encrypted communications (i.e. Signal) by simply getting Qualcomm, TSMC and Samsung to insert Exceptional Access mechanisms into the ARM TrustZone “Secure World” OS or even hack into Qualcomm, TSMC and/or Samsung, compromise their code repositories for the “Secure World” OS and point fingers at other countries being the culprit.
Will switching to Chinese made ARM A series help ? Unlikely as the Chinese ICs would probably also have inserted their own Exceptional Access mechanisms into the “Secure World” OS within Huawei or Chinese built ARM A series chips.
Shannon • May 18, 2017 6:13 AM
“Signal is a secure messaging app with no backdoor, and no large corporate owner who can be pressured to install a backdoor.”
No large corporate owner. You are misleading your readers with semantics. Signal is owned by Open Whisper Systems, which is owned by Quiet Riddle Ventures LLC. What does the “C” in LLC stand for, again?
nobnop • May 18, 2017 6:27 AM
Security, but nor privacy, see https://www.jwz.org/blog/2017/03/signal-leaks-your-phone-number-to-everyone-in-your-contacts/
Little Bird • May 18, 2017 6:34 AM
Suddenly, Congress figures out all the laws they passed to spy on everyone, literally, all the time, apply to them, too.
I suspect NSA/FBI could jail at least half of Congress right now if they wished.
I am not sure assuming Congress has the power (or will) to end the Crypto War is justified. However, their adversaries are powerful, determined and mean.
Bottom line: It ain’t over yet.
Rufo Guerreschi • May 18, 2017 7:12 AM
I am very disappointed that you suggest this is a victory, without outlining how very limited signal, eveb if perfect, would be in defending attacks by innumerable actors through malware in stack beyong applications, including OS, CPU and fabrication.
You are actually driving these users to trust the use of Signal much more than it is good for.
Jared Hall • May 18, 2017 7:28 AM
@Bruce: Yes. I think the government has taken the approach that if it wants to conduct blanket inspection (aka. Prism and other methods), the rank and file can use encryption to protect themselves. Very much like using an old analog Wireless household phones (pre-DECT). It was never against the law to monitor your neighbor’s phone call, It WAS against the law to divulge the contents of those conversations. I hope this will eventually lead to the repeal of some of the “Super-DCMA” laws that are in exsitence today. In today’s world, this is a case of WARRANTLESS broad spectrum surveillance versus WARRANTED targeted surveillance. The way I figure it, if you’re going to put something out there, don’t be surprised if somebody picks it up. No amount of laws passed is going to stop that from happening. So, yes, I think this is a very positive step forward.
@Congress Finally Understands Privacy: Regarding WhatsApp, AT PRESENT, there’s nothing wrong with it. But WhatsApp’s technology was developed by Moxie at Signal. Considering how fluid security is, it might be a good idea to go with the source. As people also point out here on this forum, the “devil is in the details”. Or, in this case, the implementation. That seems to be more aligned with Moxie’s strengths. With the contanimation from the idi0ts at Facebook, I think they will mess up WhatsApp in the implementation of new features, SOONER than later. That’s my $0.02 anyway.
Mike • May 18, 2017 7:31 AM
Maybe I am just overly cynical, but I think the people in power just found another thing they feel they should be entitled that they will legislate into illegality for the commoner…
TS • May 18, 2017 7:51 AM
im seeing a number of your recent articles on: https://www.security.nl/
a few days earlier. where do you usually get your sources ? :>
Alex • May 18, 2017 8:28 AM
I’m seeing this as a very BAD thing. It just tells me the Senate doesn’t trust its own IT department, nor its government, to keep their communications private. One must wonder what the representatives of the people of America are so afraid of being divulged… Could it be the $250k cash sitting in their freezer?
Patriot COMSEC • May 18, 2017 8:45 AM
OK, I’ll say it: Signal is a joke and we all should know it.
If the end points are not secure, then its a charade. People want security to be one click, fast, and easy. It is none of those things. They want to encrypt on a device that has access to the internet. Sure, and would you like some cream and sugar too? And a kiss from Maryland? Moscow? Tel Aviv?
Although it seems heartening that a security product for everyman is out there on the market, do you think Titan would allow it to play on the platform if it actually worked in practice?
Jared Hall • May 18, 2017 9:07 AM
@who? “We have not won the Crypto War”. Yes. we have. Perhaps a better context would be “We have won the Crypto Battle”. To your point, “We have not won the Security Battle. We have not won the Privacy War”. As others have said as well here, like Bob Dylan’s Happy Face and Patriot COMSEC, exploitation of the endpoint is the “low hanging fruit” to go after. To quote @Bruce, “If the NSA wants in, they’re in”.
@milkshaken: Good points. However I disagree with your assessment of Microsoft. It is likely they DIDN’T KNOW. I responded to Microsoft’s Brad Smith with some thoughts of my own here: https://www.jaredsec.com/2017/05/16/mr-smith-goes-to-washington/
I recall an old NSA course called “Exploitation Management”. The NSA and CIA are great at managing exploits of ALL kinds (usually HUMINT-related). I wouldn’t consider either of them great at finding them ON THEIR OWN. That said, I submit to you that the name “EternalBlue” says it all. No, I smell a rat. And its name is “Big Blue” (IBM). This is right up IBM’s alley. This is something they do well. Oh, did I forget to mention, IBM wrote the original SMBv1 protocol (LANMAN). I sure hope Microsoft’s Brad Smith is on the phone with IBM’s legal council……
Clive Robinson • May 18, 2017 9:16 AM
@ Patriot COMSEC,
OK, I’ll say it: Signal is a joke and we all should know it.
Err, no we don’t know if Signal is a joke or not. It may turn out to be the best application of it’s type we have currently.
But it’s not Signal that is the problem, it’s the connected devices it’s run on with their insecure OSs and hardware controled not by the phone user but by the service provider.
Which means that at any time the service provider can download an update to the device that changes the hardware drivers the OS and application use via a “shim” attack. Such that they intercept the keys the user presses before they get to the Signal app, and intercept the output of the Signal app before it’s displayrd on the screen, and make this avialable for testing / customer support just as CarrierIQ did a few years ago.
As I’ve been saying here and in other places for more than a decade or two you need to be able to secure the actuall end point from the open communications path. I’ve previously detailed in depth what would need to be done to secure bank transactions with an extetnal token that places the human into the last step of the communications thus act as a way to prevent such shim and other low level end point attacks.
Thus the information has been around long enough for ordinary design engineers not to make such mistakes. But at the end of the day it’s the users that want a single device that is all singing and all dancing and not have to carry anything else… And as marketing drive the product design requirments, what they think the customer wants is what the engineers build, good or bad, secure or insecure, irrespective.
Hence my first comnent in this thread in reply to JG4.
Clive Robinson • May 18, 2017 9:23 AM
@ Alex,
I’m seeing this as a very BAD thing. It just tells me the Senate doesn’t trust its own IT department, nor its government, to keep their communications private.
It could be because the Senate finally realise that their IT deparyment are helpless, and likewise their Government as well due to the compleate crap nature of what’s available, and the fact the Senate don’t want to give up their toys to be secure…
albert • May 18, 2017 9:25 AM
There is still an internecine war going on within the Gov’t, and no branch is immune. This is part of the Trump Legacy. Ideally, such a war would completely dismantle the system.
Get the beer and popcorn ready; it’s been an amazing show so far…..
. .. . .. — ….
Pay no attention • May 18, 2017 9:46 AM
“Signal is a secure messaging app with no backdoor, and no large corporate owner who can be pressured to install a backdoor.”
And it’s run by monks for the benefit of mankind…
BTW who said a large corporate owner was required to compromise a project/util?
And then consider this is used on large corporate pre-compromised phones.
Are large corporate backdoors even required to subvert apps on known-shenanigan hardware/OS platforms?
and another thing! • May 18, 2017 9:53 AM
” This is no different than denying access to firearms to select members of the population. ”
Which we also don’t really do in this iteration of America..
David Leppik • May 18, 2017 10:31 AM
@Pay no attention: the more people are involved, the more pressure points there are. An LLC owned by one person is essentially a legal fig leaf over a single individual. Public corporations can be compromised in any number of ways: through an individual employee, or by threatening profitability or share prices.
Everyone else:
I think Bruce’s point is that the Senate has determined that the government can’t protect them from spying, so they have to rely on Signal rather than some NSA-developed software. They can’t say “Signal for us and our contacts, backdoored software for everyone else” because their contacts include Americans and foreigners from all walks of life. After all, what good is perfect encryption within the Senate if Anthony Weiner’s Snapchats still make headlines?
It means they have skin in the game when Apple says, “we can’t compromise this person’s iPhone without compromising all iPhones.”
Strong encryption is only one part of protection, but it’s a necessary part.
George H.H. Mitchell • May 18, 2017 10:57 AM
I just went to the Google Play store to look at the Signal app. It requests just about every permission in the book. So exactly what is the financial model of Whisper Systems? Do we really trust the app isn’t going to abuse all those permissions to the benefit of the company?
Pay no Attention • May 18, 2017 11:18 AM
“we can’t compromise this person’s iPhone without compromising all iPhones.”
What if that were just a plausible assertion for public consumption? Do we “know”?
Obviously they have something to gain by presenting a facade of impenetrability.
Opsec requires it also. Whether it’s true or not, people surely do believe it.
I’m sure the folks who have written themselves access into those systems also have a motive towards maintaining this facade, and less directly the profitability of the ‘host’, for the purposes of maintaining access to the information they seek.
Am I being cynical, absolutely. A cynic/paranoid can only be proven wrong when you’ve exhausted all possibilities. And that’s roughly the kind of effort it would take to find out for sure. Meanwhile, everyone and their grandmother is using these systems, assuming they’re completely secure. As far as anyone knows, it’s a win/win.
But it’s actually a win/foo/win.
http://www.businessinsider.com/prism-prism-spying-project-slides-2013-6
Don’t worry, your secrets are safe with them… just don’t ask for oversight.
Pay no attention • May 18, 2017 11:41 AM
A prominent example – Intel Management Engine. On every single chip bearing the name these days, in some secret flavor or another. Oh, it’s secure. Trust it. It says so.
Your “strong” encryption is strong indeed, theoretically, given a perfect implementation.
Point to one for me.
When you’ve got motives beyond the scale of economies to subvert these systems, (in secret, yes) and the ability technologically to keep these things obscured for around 1-2 decades after they’re widely used… you tell me what ‘skin in the game’ means. The flesh is what they’re concerned about.
If you could predict only 1 roll of dice out of 1000, it wouldn’t be very useful to a gambler trying to make a buck. But given infinite rolls, or knowing exactly which roll is predictable? It’s a different sport now isn’t it?
Full disclosure – I don’t have a scanning electron microscope or a lab to strip silicon. I don’t have a team of PhD’s on a secret payroll. All I have is the track record of unaccountability and subversion to point to and no convincing reason to think it’s changing soon.
Another example, quantum machines. More secure? Sure, that’s possible.
What other possibilities exist?
Pay no attention • May 18, 2017 11:51 AM
Case study : A simple app. (Not nearly as huge/complex as an OS/platform)
http://www.bbc.com/news/39960721
https://forum.handbrake.fr/viewtopic.php?f=33&t=36364
“but don’t worry because it’s signed with a certificate with nigh-unbreakable encryption.”
In a trust web you only need 1 hole.
Winter • May 18, 2017 12:10 PM
@George
“I just went to the Google Play store to look at the Signal app. It requests just about every permission in the book.”
You would like a text messaging app that sends SMS without access to phone and wireless, no access to files or contacts? It should handle voice and video conversations without access to microphone and camera?
I wonder why such an app does not exists?
JasonR • May 18, 2017 12:23 PM
@Winter – location & calendar? Those are a bit sketchy and should be optional, as should camera. Location, to find the closest server? Plenty of ways to do that without coordinates. Calendar, perhaps to save meeting info, but again, should be optional.
k15 • May 18, 2017 12:27 PM
Does anyone do pen testing on the bricks-and-mortar infrastructure that civilization relies on? Or are we all too busy with the internet?
So, encryption for me but not for thee? That doesn’t sound like winning.
ab praeceptis • May 18, 2017 1:56 PM
Rufo Guerreschi
While I agree with your statment, particularly re. the danger of Jane and Joe user blindly trusting in Bruce Schneiers advice and with a high probability being careless then …
I’m wondering about your own undertakings. After all you use a Bruce Schneier statement in a prominent location in the marketing of your own project. Looks like Bruce Schneier is a good guy when it fits you and an evil guy when he trumpets for your competition.
I find that to be in very bad style up to the point that you seem to create attention for your own project by bashing our host on his own blog. Bad style.
Being at that: Your own project tries the “perfect security” game, too. And yes, I found some quite remarkable and good approaches and elements. Unfortunately, however, I also found enough weird and doubtful points (and even some marketing bullsh*t blabla) to have serious doubts.
Ben A. • May 18, 2017 2:32 PM
@Clive Robinson
“Err, no we don’t know if Signal is a joke or not. It may turn out to be the best application of it’s type we have currently.”
See this:
A Formal Security Analysis of the Signal Messaging Protocol
“We extract from the implementation a formal description of the abstract protocol,
and define a security model which can capture the “ratcheting” key update structure. We then prove the security of Signal’s core in our model, demonstrating several standard security properties. We have found no major flaws in the design, and hope that our presentation and results can serve as a startingpoint for other analyses of this widely adopted protocol.”
ab praeceptis • May 18, 2017 3:49 PM
Ben A.
Classical problem: people hear “it was security analyzed and proven correct” and assume that to mean that the testee, in this case signal, is “bulletproof”.
Well, no, it’s not. Neither has signal been proven secure!
What has been proven (if one is ready for a rather loose definition of “proving”) is that the model the examiners abstracted is secure under quite some assumptions, quite some of which boil down to rather wanton premises such as “building blocks like e.g. DH are assumed to be secure”.
What one uses, however, on ones smartphone is not an abstracted model and assumed to be secure building blocks.
What one uses is not even the full model of signal (which btw to my knowledge has not been formally specified, nor modelled).
What one uses is an implementation that may or may not properly implement the model, that may use rather troubled building blocks, and that is to my knowledge not verified nor even prepared for verification/analysis.
And all of that mainly in java and javascript – no more questions, thank you.
Oh and, as we’re just there: batches of ephemeral pubkeys on some server is an ugly attack surface. Not even just because it translates to a time window for nsa and accomplices, also not for what comes down to some kind of DOS attack (by depleting the batch) but mainly (well at least that’s what I focus on) because considering the many practical PK implementation problems like very strong key bias, bad random, etc. I’d be a very, very happy nsa cracker and say thanks for all the material that tells me lots and lots about your system.
May others heap praise on signal, I don’t trust it any more than a rattlesnake said to be friendly and safe.
Ben A. • May 18, 2017 4:00 PM
@ab praeceptis
I’ve read and understood the report and I’m aware of the model they’ve used but I’ll say this: Signal is more secure than ordinary cellular calls made over the 2G/3G/4G network. That doesn’t mean it’s impregnable and no sensible technical person would suggest so either. Referring to the app as “secure” doesn’t mean it offers 100% security.
The SS7 vulnerability compromises ordinary calls/texts. We could argue all day about the yet to be quantified nature of mobile baseband or the whole stack but I’d still argue that Signal is more secure than an ordinary call.
Winter • May 18, 2017 4:02 PM
@ab p
“May others heap praise on signal, I don’t trust it any more than a rattlesnake said to be friendly and safe.”
So your advice is to give up? Not use cryptography? Or do you have a better suggestion?
Clive Robinson • May 18, 2017 4:16 PM
@ Ben A.,
See this: A Formal Security Analysis of the Signal Messaging Protocol
That looks at the protocol, not the application and all the other bits of software around it. As Bruce and others have noted on the odd occasion, it’s usually not the protocol that’s weak but the implementation.
Further you can prove an algorithm or protocol secure under a set of assumptions. However as we know AES passed that sort of analysis but was found quite weak to side channel attacks that leaked key info in the test code that was downloadable and ended up in various libraries. As a result there are still products out there with side channel issues.
It will take a lot of effort to go through each and every part of the Signal app to check it’s security against “Known Knowns” and some “Unknown Knowns”, but it will be pure luck to catch any “Unknown Unknowns”.
But as I’ve said we already know that the environments Signal is designed to run in, is by no means secure and can be subverted a number of ways at a number of levels including below the CPU level.
ab praeceptis • May 18, 2017 4:31 PM
@Ben A.
“signal … more secure than ordinary cellular calls made …” – If you define “secure” as “more secure than taking in cyanide while touching a high voltage wire and standing in boiling radioactive waste” then yes.
That comparison is nonsense. normal mobile com and signal are completely different things, the latter one expressly promising security the former not.
I agree that secure probably can’t mean “100% secure”, alone for the fact the we humans who design and build stuff are, well, humans.
I would, however, expect reasonable and solid attempt as a very minimum – which signal does not provide. In my minds eye signal is an example of “less utterly mindless and crappy than ratsh*t” hyped up to “sekkure!!1!”
@Winter
No. My advice is to finally undertake reasonable attempts to create some safe and secure implementations.
We have grave limits, intellectually, practically, and otherwise (e.g. lousy OSs), yes. But we also have quite some experience, understanding, and importantly tools to start doing better.
doomed • May 18, 2017 5:15 PM
“Does anyone do pen testing on the bricks-and-mortar infrastructure that civilization relies on? Or are we all too busy with the internet?”
Yes and yes.
Ben A. • May 18, 2017 5:43 PM
@ab praeceptis
“I would, however, expect reasonable and solid attempt as a very minimum – which signal does not provide.”
Your average user isn’t going to go to the time, effort, expense and aggravation in setting up a secure messaging system. They just don’t care – and that’s before you consider the learning curve and their ability.
Signal can be used by your mom – you don’t need any technical knowledge, inclination or education.
We don’t have off-the-shelf consumer products that enable secure communications and as @Winter has said, what else can we do… give up?
Using Signal is better than making a call over 2G/3G/4G for the majority of people. Those who need truly secure communications shouldn’t be using mobile telephones.
Clive Robinson • May 18, 2017 7:04 PM
@ ab praeceptis, Winter, Ben A.,
We have grave limits, intellectually, practically, and otherwise (e.g. lousy OSs), yes. But we also have quite some experience, understanding, and importantly tools to start doing better.
The problem as I saw it some years ago is that we have no way to verify in a meaningful way the systems we use are actually secure.
If you look at the computing stack there is only so far down we can go on our own skills. We eventually reach a point beyond which we can not go and we have to blindly / foolishly trust others over whom we have no control or ability to verify their behaviour.
Thus it does not matter what tools or languages we use at higher levels, because we can neither directly see or control the lower levels. Thus as has much more recently been pointed out the likes of Intel / AMD / ARM have incorporated black box hardware that can take control of the main CPU at any time, change memory locations at any time thus control the state by which our programs work in ways we would not wish.
Do we just through our hands in the air and blindly accept that we can not trist the computers we are at the point of being utterly dependent upon? At first sight it would appear that that is now our fate. However with some extra thought there are ways you can “mitigate” most if not all of the problems.
So I started researching and building prototypes etc. After I had got things beyond a certain point I started talking about it on this blog. Some disliked, some liked but most were not realy interested.
The point is there are ways to redesign hardware to detect the low level attacks from all but those who have physical access to the system, and we need to start working out how to get them built.
@Thoth has adopted a subset of the ideas and it will be interesting to see the product he is currently talking about when it gets to the point of being ready to put on the market.
@Clive Robinson
“@Thoth has adopted a subset of the ideas and it will be interesting to see the product he is currently talking about when it gets to the point of being ready to put on the market.”
It is currently in the market at this moment. It is in it’s infancy (V1.0) and I am already planning out V2.0.
The goal for V2.0 is to incorporate a VM-in-VM system (Dynamic Secure Execution Environment) and due to a single DSEE environment running is rather slow, it will leverage the same design as V1.0 to cluster a whole bunch of DSEE capabnle V2.0 hardware if you have enough USB ports) and finally when the time comes to create a dedicated hardware to avoid the use of all those USB ports, that will take a while as I build up my team.
If you noticed, the current V1.0 is just the initial run to see the viability of such hardware implementation of clustering a whole bunch of smart card chips (in USB form factor) to turn it into a coherent system whereby each smart card chip may not be very fast but due to the clustering algorithm, every additional hardware to the cluster will improve the speed. A single hardware in V1.0 is capable of 80 encrypted password verifications per minute but every other hardware added, it increase the speed exponentially.
Also V2.0 will feature a new and improve protocol which will be much more efficient than V1.0 which is slated to make every hardware run at least 120 verifications per minute and then multiply exponentially by every hardware added.
The goal is that these hardware are still far from the Prison design and hopefully one of the future versions (if all these last long enough :)) and also with enough resource, to create “mediator” chips as per described in your Prison architecture.
I guess I owed you a couple of barrels of beer for attempting to implement your designs ?
Link: https://thothtrust.com/products/securipass/docs/SecuriPass%20Brochure.pdf
ab praeceptis • May 18, 2017 8:07 PM
Clive Robinson
Difficult issue, really difficult. One might, indeed, feel/think that any efforts are in vain anyway as some other layers are unsafe/insecure anyway.
I personally tend to approach from the other perspective. The main reason being that while nsa and accomplices are abusers and sometimes “sponsors” they are rarely the creators and moreover most creators, so I guess based on my experience, are not evil.
It is, in other words, the very properties of the problem domain that provide reasons for hope and cures. Asking why means asking for a realistic view and the reasons. Here’s my take:
Education, arrogance hand in hand with ignorance, a sizable bag of laziness, and lack of tools. In that order.
And the ever turning profit greed wheel considerably adding to all of the factors.
Thanks to Snowden and plenty of scandals more and more people understand that we must change IT development – unfortunately most of those people do not see/know ways how to do better (which to a large degree is due to bad education and ignorance).
Plus: It’s still hard, it’s still lightyears away from firing up ones IDE and having some buttons to comfortably click.
Let’s start with spec and modelling. Most do not even grasp the concept and think maybe (if that) of uml (yet another cancer in the xml plague familiy). Not even knowing what it is and what it’s about is not a good point to start at …
Next, the tools. The B or tla[+] tools are not exactly evident and easy to use, particularly for people lacking the necessary knowledge.
Moreover, one needs quite some experience (which is quite rare).
Looking a programming languages major social factors enter the game. Humans tend to go with the mainstream which is C, java (plus a plethora of interpreted languages) – all of them having hardly wasted a thought at safety. To make it worse, usually the most widespread languages also have the most comfortable IDEs, editor support, literature, articles, etc. And, of course money; why would I learn and get expertise in some “weirdo” language when javascript promises to easily get a job and easily earn a life?
I’d like to close with a positive sparkle: If we chose and walk the right way we will have nice leverage effects on our side. Solid safe software, for instance, will support us in better testing hardware and better hardware will be a better basis to run our software on.
We today can create solid and safe software. It’s not yet a comfortable way to walk and we will are only a few in the beginning but the way is there. Let’s walk it!
Ex-Pat Mark • May 18, 2017 11:47 PM
Well, the members of Congress also get excellent health insurance….
Patriot COMSEC • May 19, 2017 12:01 AM
@ Clive Robinson
I agree with you. I was trying to highlight the fact that Signal has been designed for use on unsecure, and unsecurable, systems. Therefore, it is a joke, and the joke is on the users. Any other belief ignores facts. People are being told a simple story about security, and what they are not thinking about is the urgent and immense national security interests of several countries, and the huge rewards for others, pushing for the ability to subvert Signal and collect data.
Other pertinent facts include the recent disclosures about how the iPhone is wonderful for being controlled remotely, and how Android’s apps make it exploitable.
Yes, Signal has done impressive work, they have been repeatedly recommended by Snowden, and we even hear, from illegally-disclosed NSA documents, that the NSA regarded Signal as a major threat in 2012. From those same documents we learned that TAILS, TOR, and TrueCrypt were regarded as even more dangerous, as catastrophic. So why the difference in threat level? What is the difference between “major threat” and “catastrophic”? Isn’t it reasonable to guess that the difference is between subvertible and we-can’t-own-it? In other words, if it were an inaccessible system to the U.S. in its actual employment, I think we would be hearing the FBI scream.
As soon as one sees the Google and Apple logos on the Signal website, that should be enough to scare us. The systems of Signal + device are going to be major collection targets because of all the diplomats and other government people, all over the world, who prefer to use Apple products, and Android will be number two because of its spread.
Matt Green is excited about Signal’s code, and that is nice. The only problem is that this is just part of the game, as you know. The question is this: does the end-user have a reasonable expectation of real privacy when using Signal? I am arguing that the answer is no: the end points must be secure too.
Dirk Praet • May 19, 2017 4:40 AM
@ Patriot COMSEC, @ Clive, @ Ben A
Therefore, it is a joke, and the joke is on the users.
No it isn’t. Until such a time that @Thoth, @Markus Ottela, @Figureitout or @Clive comes up with a hand-held, affordable and commercially viable solution that physically separates the encryption and communication device, Signal is probably the least insecure communication method of its class as compared to Skype and a number of other alternatives. So it’s exactly as Susan Landau says: a step in the right direction. Nothing more, nothing less. (Susan, BTW, is always right)
For what it’s worth, I don’t think the official seal of approval by Frank Larkin’s office is going to change much on the shop floor since everyone in both Congress and the White House with even a remote clue about secure communications was already using Signal anyway. As technologically illiterate as the average senator may be, most of them probably have one or more staffers that are not and that had already previously pointed to Signal as a means to raise the bar for eavesdroppers.
@ Thoth
It is in its infancy (V1.0) and I am already planning out V2.0.
Cool. From an ergonomic perspective, the ideal device IMHO should be some sort of two-part flip device (e.g. Motorola RAZR), one of which holds a type of smart card/key for the encryption, the other part doing the actual sending and receiving. I guess there’s different ways to make them talk to each other in a secure way (miniaturized data diode/pump?)
@ Alex
It just tells me the Senate doesn’t trust its own IT department, nor its government, to keep their communications private.
Either at some point you wake up and smell the napalm, or the napalm puts you to sleep forever.
herrman • May 19, 2017 6:13 AM
Maybe I’m being optimistic, but I think we just won the Crypto War.
Well I don’t share your optimism here. Governments around the globe are used to different rules for “them” and for the “common folk”.
JG4 • May 19, 2017 6:25 AM
@Thoth, Clive, Markus and others who have helped
Thanks for your diligent efforts to teach and implement the important facts.
More fascinating news than most people have time to read.
Links 5/19/17 | naked capitalism – Tor Browser
http://www.nakedcapitalism.com/2017/05/links-51917.html
…
Julian Assange: Sweden drops investigation against Wikileaks founder based in Ecuador’s London embassy Independent. Deck: “Scotland Yard says it will still arrest Mr Assange over skipping bail if he leaves embassy.” So…
http://www.independent.co.uk/news/world/europe/julian-assange-sweden-drop-charges-wikileaks-ecuador-embassy-london-sexaul-assault-rape-us-a7744181.html
…
Cage director risks prison over refusal to disclose password to police Middle East Eye. Trouble at the border.
http://www.middleeasteye.net/news/cage-director-faces-jail-over-refusal-disclose-password-during-airport-interrogation-681177289
US looks at extending laptop ban to all flights FT
https://www.ft.com/content/a5624c3c-3bd7-11e7-821a-6027b8a20f23
Facebook to Vestager: Let’s be friends Politico. Be sure to come back to read Maciej Cegłowski’s post on this topic later today.
http://www.politico.eu/article/facebooks-vestager-playbook-lets-be-friends/
Why Hardware Engineers Have to Think Like Cybercriminals, and Why Engineers Are Easy to Fool IEEE Spectrum (Chuck L).
http://spectrum.ieee.org/view-from-the-valley/computing/embedded-systems/why-hardware-engineers-have-to-think-like-cybercriminals-and-why-engineers-are-easy-to-fool
The need for urgent collective action to keep people safe online: Lessons from last week’s cyberattackMicrosoft on the Issues
https://blogs.microsoft.com/on-the-issues/2017/05/14/need-urgent-collective-action-keep-people-safe-online-lessons-last-weeks-cyberattack/
…
China?
Stratfor explains how China’s Belt and Roads Initiative might reshape Europe Fabius Maximus (Re Silc). Re Silc: “We will fund groups to kill it and blow it up.”
http://thehill.com/blogs/pundits-blog/the-administration/334105-latest-trump-russia-report-still-lacks-smoking-gun-of
…
Big Brother Is Watching You Watch
Aadhaar and an Omnipresent State That Will Never Forget You The Wire (J-LS). Yikes.
https://thewire.in/136521/aadhaar-omnipresent-state/
Uber Doesn’t Want You to See This Document About Its Vast Data Surveillance System Gizmodo
http://gizmodo.com/uber-doesn-t-want-you-to-see-this-document-about-its-va-1795151637
@Dirk Praet
“the ideal device IMHO should be some sort of two-part flip device (e.g. Motorola RAZR), one of which holds a type of smart card/key for the encryption, the other part doing the actual sending and receiving. I guess there’s different ways to make them talk to each other in a secure way (miniaturized data diode/pump?)”
It really depends on how much security you want. If you are using a chip you are comfortable with and also with sufficient trust you have in the hardware, you can just do a smart card with a single input/display controller (i.e. embedded OLED screen with button controls).
If you are paranoid enough, the entire data diode concept have to come in and if that is the case, this wouldn’t be in a mobile security form factor but more of a secure compartmentalized workstation with all the added security measures including suitable physical security and EMSEC and all that stuff.
Patriot COMSEC • May 19, 2017 9:02 AM
@ Dirk Praet
-hand-held
-affordable
-commercially viable
There is one more prerequisite:
“We didn’t have to do anything.”
(Look on the Open Whisper System website and read the text on the phone)
I did not have to do anything. That is voice of their average user.
I am not putting down the work of those capable, clever people who invented Signal. My hyperbole was aimed at this ugly ditch between the strength of the math and the weakness of the application.
We need to make a solution that is easy for everyman, that marries the strength of the math with something requiring key and lock surreptitious entry to subvert it.
Dirk Praet • May 19, 2017 11:00 AM
@ Patriot COMSEC
We need to make a solution that is easy for everyman, that marries the strength of the math with something requiring key and lock surreptitious entry to subvert it.
There’s no such thing. If you want some decent level of security, you are going to have to work for it yourself too. Consider this: at some point an entire generation of musicians decided it was no longer necessary to properly learn how to play an instrument either. We’ve been stuck with this horse manure called techno ever since.
V.R.K • May 19, 2017 12:56 PM
great dialog folks. thanks much Bruce. Peace to this house.
only…
Tell the guys outside my window who are running daily refinements on side channels that “least insecure” is helpful to my cause. Tell even the dozen or so “in striking range” trusting users of active generic devices in my world who aren’t using my “super-duper secured device”. Tell even my landlord who does hard resets on my best efforts when I’m not home. Tell my vigilante neighbor upstairs who freely lowers devices at will. Tell the power company who runs the “smart meter” here…
Build your own “best case”, people.
Infra red laser and glass fiber comes to mind for the office.
Or use Gandhi’s solution: announce your secrets.
Even in SILENCE your secrets will be KNOWN: Your privacy is a whore on her back for tax payers, and spin fodder for insatiable lying baiters.
Im out.
…- .. –. .. .-.. .- -. – . .-. — .- -.. -.- .. .-.. .-..
ha ha ha haaaaa…
Figureitout • May 19, 2017 10:50 PM
Moxie never said Signal’s intent is to stop targeted surveillance, it’s to stop blanket surveillance. Some of the measures we lay out here (I try to implement them and provide guides) are for that, but almost no one wants to put in the work, which is understandable when we’re all overloaded w/ work and other stuff just to survive. Plus the people you want to contact most will screw it up anyway.
The fanboi’s proclaiming it to be secure from targeted surveillance need to be challenged too b/c they’re wrong, but putting it down otherwise is just jealousy imo. Any successful product will get hated on and attacked from all angles. If you hate it, the best way to hate is to hack (in legal manner).
Dirk Praet
–If you want worldwide comms that don’t rely on internet or cellular infrastructure, a handheld device may be pushing it if you want reliable error-corrected transfer. It can be done (just going to use more energy than being efficient and trusting in a world of angels), just a matter of the demand being there, we need more hacking of computers, people’s data and infrastructure to get that demand. That’s how I got into computer security, getting owned one too many times.
Being able to separate the encryption and communication, that should definitely be doable. Any architecture I would build as of now would have a lot of serial data diodes w/ optoisolators transporting something from A to B, so long as no return path is needed.
As far as what I’m working on (security-wise, got another huge project and work…I’m going to have a hard RF project on my hands soon, don’t want more hard comms projects! :p). I’ve got reading contents of a file off a SD card to a serial port on an arduino (just a single wire for transfer, beautiful simplicity; and bet if I ran the test transfer a million times, it’d pass 95% or greater). Adding the features I really want is going to take some late nights…Transferring the actual file over the serial port will be tricky (what I’m thinking is to first get the file name, send that over, and create a file w/ that name on receiver, then start sending contents but it needs to be raw contents, then close it). First step is just sending one full file, then do file diff. Next step is doing this for an SD card full of files (text, source files, PDF’s, html, docx, etc.).
What you have then is one of the strongest methods I know of to transfer files from airgapped PC’s. As always, hacks or attacks (not worthless blabbering) on it welcome.
Rachel • May 20, 2017 6:33 AM
@ Dirk Praet
‘Consider this: at some point an entire generation of musicians decided it was no longer necessary to properly learn how to play an instrument either. We’ve been stuck with this horse manure called techno ever since. ‘
nice analogy, without derailing OT Two examples come to mind we can compare precisely with infosec.
Firstly the entire generation of musicians in the ninties, upon the advent of digital recording methods and Pro Tools, decided ‘hey I don’t need to practise anymore! I can fix any mistakes in post production! And why use our drummer in the studio when we can use a drum machine! No need to know how to sing- there’s auto tune!’. heaps of albums experienced all of these ugly catastrophes
The other analogy, is the original electronic music that was analog ,eg Kraftwerk amongst others. created by genuine innovators with real ability,even Steve Albini is a fan of it. All now completely ignored, disregarded and unknown. All thats known as electronic music now is ubiquitous, not requiring much in the way of skill or musicianship, and extremely popular.
Dirk Praet • May 20, 2017 8:19 AM
@ Rachel
The other analogy, is the original electronic music that was analog ,eg Kraftwerk amongst others.
Funny you should mention Kraftwerk. They’re playing my home town tonight and there’s a fair chance I’ll drop by there if I sufficiently recover from my yesterday night’s hangover. I’m kinda curious if afterwards they’ll also show up at the W-P bar round the corner like Bob Dylan did last month. Everyone was too gobsmacked to even think about selfies or autographs and respectfully left the man alone.
@ ab praeceptis
Yeah, yeah, we know.
@ Figureitout
What you have then is one of the strongest methods I know of to transfer files from airgapped PC’s.
I concur. That’s why I think it would be kinda cool if such a setup could be miniaturized to fit a two-part encryption/communication handheld.
stranger_around_here • May 20, 2017 10:52 AM
Somebody probably mentioned this, but . . . this is the same Senate that, uh, has excellent health insurance and would deny the rest of us the same. And I’m supposed to believe that they will respect our right to privacy?!
I’m not buying it. (that’s a period)
ab praeceptis • May 20, 2017 3:43 PM
Dirk Praet
Apologies. My only excuse is that I’m quite concerned about IT security, so I dared to comment on something related to and important for IT security on this music blog.
P.S. I like Kraftwerk, too, and actually helped them a tiny bit when they started. But failing to see a relation to IT security and lacking boundless egocentricity I considered that private and irrelevant here.
Again, apologies for disturbing a music blog with security stuff.
RedPhone and GCM are closed source and the contact list isn’t private.. LibreSignal was blacklisted for reasons people still don’t understand.
Futility • May 22, 2017 6:17 AM
I also doubt that this signifies the end of the crypto war (as detailed by other commenters already). But what worries me even more is that the linked article treats as established fact that the Russians meddled with the 2016 election when it’s anything but. All we have seen so far are indications of Russian involvement which could just as well have been deliberately planted by the real perpetrator. But by endlessly and uncritically repeating it, it becomes an accepted part of the narrative – what everybody just knows -, playing into the hands of the Democratic party establishment that is only too eager to deflect attention from their colossal failure in the 2016 election.
splif • June 3, 2017 4:42 AM
it’s a good thing, any layer that can be applied without causing too much strife
for them is a good thing, all it is, another layer
Subscribe to comments on this entry
Sidebar photo of Bruce Schneier by Joe MacInnis.
Ben A. • May 17, 2017 2:56 PM
Maybe I’m being optimistic, but I think we just won the Crypto War.
Ever the optimist @Bruce. I don’t think we’ve won the Crypto War per se because, for me, the Crypto War is about governments trying to restrict strong cryptography. We don’t have any assurances that they’re not going to try and pull a fast one.
Maybe the new Trump Executive Order which “promises to hold agency heads responsible for slipups” may explain this volte-face.
http://www.theregister.co.uk/2017/05/11/trump_cybersecurity_exec_order/
Some interesting commentary here:
https://news.ycombinator.com/item?id=14354177
Back to being pessimistic:
“Any Half-Decent Hacker Could Break Into Mar-a-Lago”
https://www.propublica.org/article/any-half-decent-hacker-could-break-into-mar-a-lago