Schneier on Security

David • November 7, 2017 9:45 AM

@Dan H: As has been said by others, depending on where the breach has occurred the signature can also be compramised. But even if it hasn’t been, in a long exchange of emails, how likely is it that people are going to check for every single message that the signature is correct?

Fundamentally for security to be good it must be easy and automatic, and email is far too clumsy and old to be made workably secure. Even experts struggle to use ‘secure’ encrypted signed email and even then the emails are frequently all encrypted with the same key that changes infrequently if ever, so if it’s ever lost/broken all the messages ever sent become readable. [assuming they were intercepted in the meantime, and/or still available on the mail server]

This means that email is fundamentally and unfixably insecure, there is effectively no such thing as secure email and it should not be used for any communication purpose where security/privacy is required.

More secure communication systems [such as Signal, Whats App, etc] that use end-to-end encryption with keys changed every time you communicate should be used for communications that require privacy and security…. which is pretty much all communication… We should get rid of email entirely. Which will of course require some redesign of systems that assume people have email addresses for signing up to things, and use something else [or a selection of somethings else] instead. Which will have the added benefit of not having peoples insecure unencrypted emails as the key to all their other accounts for everything else they use on the internet.

The problem, for people like me, is that Signal, and Whats app et al, were designed for smartphones. I do not own a smartphone, and never will, and thus cannot access these secure systems as people do not seem to want to create such similar systems that work for desktop/laptop pc’s without requiring at least a smartphone account to set it up as Signal does. And I am not prepared to pay to own a personal tracking and recording device specifically so I can have a secure private communication program so I don’t have to be tracked or recorded. This assumption that everyone owns, or will own, a smartphone is deeply misguided.

David • November 8, 2017 4:21 PM

@Clive Robinson: You say that “Such applications** in use ARE IN NO WAY SECURE nor can they be.” however I respectfully suggest that you have missed the point.

Perfect security is impossible, thus demands that something have perfect security or be dismissed as “in no way secure” are thus demanding the impossible.
In the same way that as it’s impossible to be absolutely certain about anything some claim it’s thus impossible to know anything, which is absurd. Knowledge is simply uncertain and comes with a probability attached to it.
Similarly security is not perfect and comes in degrees; things are more or less secure, and also more or less useable. [Frequently but not always as a result of how secure they are. Email as a case in point becomes less and less useable the more secure you try to make it, rapidly becoming practically unusable long before it becomes usefully secure.]

Breaching security has a cost, and is done to get some benefit.
If the benefit is greater than the cost an attacker stands to gain by breaching my security.
Thus I do not need perfect security, what I need is [in general] security that makes the cost of access higher than the expected benefit.

Email does not achieve this [clearly] and thus can be labelled as deeply insecure, the cost of breaching the security is so low that it’s almost free and mass surveillance of everyone is possible and indeed done multiple times over.

Programs such as Signal and Whats App et al do [in general] achieve this, because the cost of breaching the security is higher than the expected benefit for most attackers for most people and they prevent mass surveillance of the messages being sent. [With varying degrees of privacy]

It is still possible to break into the various devices [PC, Mac, Smartphone, etc] that they might run on and get at the data but there are steps that can be taken to make that hard enough that generally someone has to really want to attack you and get at your data to apply enough resources to get in. [The previous story on this blog being a case in point]

As for using the apps signalling that you have something to hide… There are over 1 Billion Whats App users… There is a safety in numbers. The more people use these secure apps for their communications and the more normal and widespread they become the more resources it would take to break into everyone’s device to bypass there security and read their messages. At a Billion+ users Whats App long surpassed the numbers where that was practical for anyone.
Using these services isn’t flagging you up as having something to hide; it’s simply becoming the way we communicate.
I would just like that functionality to be available on devices other than smartphones. For those of us who will never ever use them.

[**Apparently referring to ‘Signal and Whats App etc etc’…]

@me: I’m generally agreeing with you but… How many people do you think could actually use a “Qubes OS disposable VM”..? I volunteer to help people learn basic computer skills and there are still plenty of people out there who are still struggling with the basics of how to use a mouse/keyboard, how to access emails/the internet, and what a web browser is… etc. And your suggestion is that to avoid fraud while they are going through the stress and hassle of trying to find a house they should set up a VM to make sure their email is more secure? If those are the lengths you need to go to to fix email, email is so broken it needs to be dropped. And replaced with a system that regular people can actually use. One I can explain to someone who is 70-90 years old and has never used a computer/smartphone/tablet before but is now required to use one to access certain government or other services.

David • November 12, 2017 6:06 AM

@Clive Robinson: “The commercial consumer OS developers have done it for the sorts of attacker you need to worry about to protect themselves. They take your plaintext and put it up in their cloud or ET Phone Home with it to their mothership via “Telemetry” or “Test Harnesses” to provide “Help desk support” etc etc”

Wow. Ok, well I don’t like windows unturnoffable telemetry as much as the next guy… Which is why I turned it off, the service’s are disabled and the relevant servers are blocked by firewalls. [which is pretty trivial to do].
But let’s say that like most people I haven’t done that…

It’s certainly possible/probable that as part of the telemetry data some of my private documents/messages could get uploaded to Microsoft/[OS provider of choice].
which is why the telemetry thing is such a big issue.
But what you are suggesting is that the security of signal et al is worthless because ALL of my messages/documents are available for free to anyone because they are uploaded by telemetry… Which is b*****t.
Telemetry doesn’t upload anywhere close to that much data [and in my case it’s uploading none, which as I say was easy enough to achieve]. People would have noticed and screamed from the rooftops if telemetry was uploading the entire freaking contents of your hard drive to Microsoft [or whoever].
So the threat of telemetry is that it might upload revealing scraps of personal data, which again is a big deal, but not the all encompassing threat that makes not encrypting and/or obfuscating the data in transit not worthwhile.

“As @Bruce and many others myself included, have repeatedly pointed out year after year it matters not one jot how strong the strongest link in the security chain is, it’s how weak the weakest link is that matters where security is concerned”

And I notice that @Bruce recommends people use Signal…

Yes, it matters how weak the weakest link in the security chain is.

However for our purposes, email is like sending messages by postcard, just about any person who cares to can intercept and read/copy/alter/duplicate/imitate your messages along the way.
It’s more secure to send messages by fax [in a fictional world without wire tapping] even if you keep all the printed out documents in unlocked draws and never lock your house because your threat exposure is still less than the people communicating via email, especially when there are a billion+ of you doing it. You can’t use the argument that all those billion+ people are suspicious criminals with stuff to hide nobody will buy that. With those kind of numbers you are just people.

Making secure apps more widely available and easier to use so more people use them for regular communication means the communications that need to be secure get hidden in everyday fluff that is encrypted to the same standard. That way the ‘bad guys’ don’t know what to target.

Speaking of which…

“As I’ve repeatedly said the sort of attackers you need to wory about care not one jot how secure the path is between the secure apps or how secure the apps are, they realy realy do not care.”

The people who were taken in by this scam apparently had to worry about the kind of attackers who apparently do care about how secure the path is between the two people communicating.
If these people had been communicating by Signal or Whats App then the attackers would have had a vastly harder time intercepting the communications undetected or sending the fake messages undetected. Hacking an email account is easy, hacking into Signal or Whats App is really hard.
This story would likely not have happened in a world where people communicated via secure apps instead of email. [criminals will of course find other ways of targeting people there will be displacement but that’s not an excuse for not fixing weak security]

So while the secure apps might not mitigate against every threat, and they might not mitigate against the threats you’re paranoid about, they do mitigate against some threats! Threats that matter in real ways to real people.

As someone who has met and dealt with people who have lost tens of thousands of pounds to ‘Microsoft scammers’ these kind of threats are very real and worth mitigating against. Even if they are not as sexy as your paranoid overarching government conspiracy new world order.

And yes, the fact that the business model of the internet is surveillance and that governments and security agencies have got addicted to that is a big deal. But it is not the only deal.
“Go back and look at the CarrierIQ debacle a few years ago to see just how easy security is to bypass when the Comms End Point reaches further than the Security End Point.”

I did go and have a look at this and this makes my point perfectly. It indeed be a choice whether or not you upload your encryption key to the cloud… but to argue that an unencrypted computer is less secure than one encrypted by that system is to fail to understand maths.
Yes there are ways that attackers could potentially get the key but it is an extra step they have to take to get at your system, and extra bit of difficulty [however small you think it is] and that makes it harder and thus more secure. The number of people who can attack that system is smaller and thus that system is more secure.
Most people are [rightly or wrongly] more concerned about criminals stealing their data/identity/money than they are about the NSA spying on them.
And the measures we are talking about definitely help protect against criminals, even if you think they are less help against the NSA.
And sure, there are probably some elite crackers who could go though these protections like they weren’t there… but the overwhelming majority of criminals that most people are likely to be ‘attacked’ by are not those elites. They are the ‘Microsoft Scammers’ of the world, the script kiddies, the scum of the Earth, with little skill and fewer morals who are after a fast buck and a little security really is all it takes to put them off and make the difference.
So you telling people that the protections these pieces of software offer are worthless because they don’t fully protect against elite hackers or the NSA … the attackers people need to ‘worry about’ is bullshit… Because most people most of the time do not actually need to worry about the NSA*.
They DO however need to worry about cybercrime.