Schneier on Security

Clive Robinson • April 7, 2021 7:55 PM

@ Weather, ALL,

They might be trying to increase the uptake of signal

The Internet has fairly well proved it’s self to be a “First to Market is the winner takes all of the market” despite what a course in economics will imply (due to the fact of the hidden assumption in economics of “distance costs”).

So from those behind it there is probably the thinking of “Signal has to grow or die”. And if it’s the latter all that they have sofar invested will be lost, or not reach the potential they are after. But tying themselves to a highly volatile trading/speculating crypto-currency is actually going to make Signal’s demise more likely.

there is apple pay that uses the phones RFID to pay for things.

But there is a big difference between a conventional fiat currency independent transaction system and a crypto-currency payment and reconciliation system and that’s “Risk” with a very large “R”.

By and large all crypto-currency systems are designed to be inherently unstable thus they tend to hyper inflation/deflation cycles as people with sufficient funds manipulate the system creating volitility to their advantage. The volitility in turn means that exchange and reconciliation systems carry bigger risk, therefore will charge significantly higher fees.

You have to only try and live in a country with hyper-inflation for a very short time to know what that does to a civil society. Then realise it has a reinforcing effect on the entire economy as more and more labour is devoted to try and shorten the time any one actually holds currency in their hand.

In fact providing a mechanism to speed up the aquire/divest cycle just facilitates hyper-inflation. It’s one of the reasons certain people carry out “High Frequency Trading” and will thus pay hansomly to have the shortest network path between them and the point of trade (yes even traders have to obay the laws of physics with respect to the delay caused by the speed of light).

So irrespective of if it’s Signal or some other secure messaging app making the movment of crypto-currency fast is not going to end well for the security of the app or it’s greater environment.

The reason is “cryptography-costs” in CPU cycles so both in work/power and delay/time. On a per bit basis the fastest form of cryptography is the stream cipher as the work of generating the key stream can be done well in advance and in most cases in parallel.

The problem with security is that it’s rather more than just hiding information via cryptography. So whilst you can hide information on a bit by bit basis other things have to be taken into account (CIA triade for old timers). For instance bitwise encryption is easily subject to active “bit flipping” attacks, because the cipher alphabet size is considerably smaller than the plaintext alphabet size. This has consequences at higher levels in the security stack especially where known message formats are used, especially when optomised to remove redundancy (ie made more efficient).

Thus with financial transactions and a highly volatile trading currency there will be a pull in two different directions. Firstly that of privacy from security which is what would be desired by ordinary users (otherwise why use a secure messaging app?[1]) which increases delay. Secondly to minimise transaction times/delay to reduce the effects of trading currency volitility which generally will act against security thus privacy.

We can see from the state of the current finance markets who will win in that game, and it will not be privacy.

But aside from the fact that security will decrease in time to reduce delay thus trading currency volitility effects, there is the ROI on “breaking the system” to consider.

At the moment very few have the resources or desire to break the security of secure messaging apps. That is the value of the messages are by and large of no value except to,

1, SigInt agencies of various forms (mostly funded by a nations wealth).
2, Information brokers, who mostly deal with bulk or industrial scale data collection.

As these are few and far between it is entirely possible that the privacy via the security has already been broken “by design” but is generally not known (Skype?).

Making messages have direct value as in financial transactions obviously attracts two other types of attacker,

3, Those looking to greatly polish their C.V.’s reputation.
4, Criminals.

The fact that they have succeeded will not be kept secret for very long at all thus there will be a significantly increased preasure to find more and more breaks in the security of the secure message app system and thus the increased loss of privacy.

There are other reasons not to get involved with high volitility trading currencies let alone “build them in” to your secure messaging app not least of which is legislation and regulation. But also and perhsps more importantly you tie the fortunes of your secure messaging app to the fortunes of the highly volatile trading currency, which with crypto-currencies is in effect a “Suicide Pact”.

It’s why they should be developing a currency independent transaction system not a currency dependent one. And that way leave all the other issues to those that take on the currency volitility risk such as financial exchanges, banks, and sovereign nations.

[1] There appears to be three basic reasons why a security messaging app gets increased usage,

1, It’s hip/trendy/cool.
2, It offers increased privacy or functionality over it’s competitors.
3, It’s got the most users you commonly interact with using it.

In the long term it’s the third reason for “winner takes all”.

Clive Robinson • April 10, 2021 5:04 AM

@ Tatütata,

such a move would be perilous as it provides a ready-made excuse to censors and bullies for attempting to control or shut down the channel

It’s actually an easy risk to mitigate.

As I noted above, the trick is to provide a mechanism to trancport “any” currency, not a locked in mechanism to transport only a suspect crypto-currancy.

That way you provide the equivalent of a “common carrier” significantly minimising the risks your service provides to what is managable.

But more importantly you avoid all the risks of association or identification with a questionable currency. You also avoid all that banking, lending, securities, money laundering, etc legislation that very few people realise how easy they are to transgress thus give not just an excuse to take action but a method to smash down hard on all involved.

It is very very unwise to provide “stored value” systems even banks etc get it wrong, but they have the reserves and stature to ensure they get only fines and the fines are managable (except in Iceland where jailing bankers almost became a competative sport).

There are only a handfull of tech companies with the reserves. But they have no stature in finance and are certainly not anywhere close to being a “too big to fail”. Thus they represet little risk that can be used for effective counter barganing.

Even the sociopath in charge of Facebook realised that, which is why he tried to partner with financial organisations, who having looked at the risk/reward split of the deal he was putting on the table smiled politely and declined rather definitely.

If Moxie and those behind him think they are any smarter, than I suggest they rather rapidly reevaluate their position…

No doubt that is why some people think it is actually a “pump and dump” activity similar to that the likes of Hedge Funds routinely do to tech company stocks.

I guess the question is,

“Is there sufficient cause for the proboscis of the genus Camelus that is possessed by the SEC to start not just sniffing along the hem of the skirt, but forcing it’s self intrusively via the rear entrance flap up into Signals tent?”

I guess we will have to wait and see.

Clive Robinson • April 11, 2021 5:01 PM

@ Winter,

<

blockquote>Untill there is a fully secure communication system available to the masses,

There is and has been for atleast half a century prior to the development of the original cellular radios. I’ve repeatedly explained the process required with the simple to understand use of a secure paper and pencil cipher[1][2]

The problem is the majority of users do not want to learn to do anything more then press a button or say the word “send”.

Thus it’s the pandering to the whims of “convenience” that stops thete being as much security as you would desire, nothing else.

So people get the illusion of security not the actuality of security with these apps.

Worse though they only find this out when the “midnight knock” happens or their loved ones find the person has “disapeared” or been found “suicided” or “double tapped” in some out of sight place, or draged outside by security services and executed violently and messily as a deterent to others. All of which and more have happened in very recent times.

Security is not a “convenience” and never has been and in all probability never can be, you have to work at it. Because security needs work, in turn privacy needs work, and in turn so does liberty that rests upon them as foundations.

Benjamin Franklin nearly three centuries ago put pen to paper and wrote,

“They who can give up essential Liberty to obtain a little temporary Safety, deserve neither Liberty nor Safety.”

Understood the notion of what was required. Thus today “safety” could be replaced with “concenience”.

[1] I used as an example the One Time Pad (OTP) which is both simple to inderstand and simple to use securely and has been for nearly a century. Whilst the OTP has many failings for the incautious, they are well understood. Unlike the myriad of failings of more convenient electronic devices. That is not to say you can not use seperate electronic devices/tokens[3], but you have to understand what you are doing and that is way way more dificult with electronic devices involved (see TEMPEST and EmSec rules for Red/Black seperation for Passive EmSec and how to screen etc against Active EmSec).

[2] https://www.schneier.com/blog/archives/2020/01/mailbox_master_.html/#comment-344813

[3] See the use of optically issolated TFC hardware security system, Marcus Ottela, developed with a little assistence of this blogs “Usuall Suspects” of the time,

https://github.com/maqp/tfc/wiki

Clive Robinson • April 12, 2021 1:35 PM

@ Winter,

When I am up to a murderous nation state, e.g., Vlad the Poisoner…

It’s not just Vlad you need worry about. Most Western governments have a “Dept of Dirty tricks and wet work” tucked into their alleged Intelligence Community, to deal with those considered undesirable.

Likewise many other nations who might employ the likes of the Butcher of Riyadh to chop up journalists, or just out-source such work via ex-government military and police of Argentina, Chili, and many other nations, run by juntas, dictators, tyrants, etc, many of whom were originally trained by UK, US, Israeli, French, etc personnel for “politicaly desirable” reasons.

As others have noted, our civil society is being dragged backwards in many ways by what are in effect extreamists masquerading as conservatives, religious leaders, conspiracy theorists, or fringe politicians. That have managed to worm their way into legislatures and even the judiciary.

Thus with technology giving them “collect it all” and next to no cost storage, in effect they have a “time machine” by which to go back and select those to be used to set examples of. We’ve seen this go on in the US where there has been “The War on being not White and Right” similar in quite a few European Nations where an Orwellian tactic of creating a popularistic hate group mentality alows the striping of privacy, rights and even fair trial, in order that new more draconian powers of control can be implemented (look at what is going on in the UK for instance with BoJo’s new legislation drive, designed to make him another Vlad).

Which brings us to,

But then how to get it at its destination?

You could start with the tried tested and still fully workable old school “field-craft” I’ve mentioned it before and how it can with care be modernized.

But you do not need to go to the lengths of dead letter drops etc[1]. AS I’ve mentioned before there are ways that are as secure as the OTP that can be used in not just “Open Communications” but “Broadcast Communications” under the eye of even the most eagle eyed of censors and Intelligence personnel and will remain secure indefinitely unlike the majority of determanistic systems.

Thus such Open Broadcast Systems are only realy going to fail you by “Human failing”, either yours or the second party in a communication. That is by either party being “turned” or willfully “betraying” the other party in the communication.

There are even ways to deal with this but I’ll leave others to say what they are, they are not difficult to work out (when you know one instance in the general class).

The more important thing is to actually minimise any damage all communications can do to you or others. One way to do this is to reduce correlation of communications to other observable actions. This is a more intetesting part of “Traffic Analysis” which is pretty much all about seeing correlations or even causation in things carried out by those you observe. Or by creating correlations with other things that are effectively innocent activities.

There are also methods by which the first party in a communication can check the fidelity of the second party and there is OSInt on how to do this in various publications, which are not that difficult to obtain.

If you do read them, then you might well consider again,

But to communicate with my family, I take Signal as currently sufficient.

Like the usage of Tor the use of “secure message apps” marks you out from the herd. Whilst it may not be enough to put cross hairs on your back most prosecuters would use the fact you use it to paint you as a criminal or worse. Especially with the likes of the FBI and DoJ and other nations equivalents alread publically saying “It’s only used by terrorists and criminals” as the recent statments around EncroChat phones has shown.

In other words you already have not “The Mark of Cain” but of Diablous on you, as far as those that believe they carry the mark of Angelus…

[1] Unfortunately the likes of Amazon with it’s Ring system is making old school field craft much harder than it used to be. In effect they are another “Collect it all” operation. Though what is not clear is how long they keep it for, thus how far back their “time machine” will reach. It’s probably wise to assume that it will be atleast seven years capacity.

Clive Robinson • April 14, 2021 5:48 AM

@ SpaceLifeForm, Dirk Wemhomer, JonKnowsNothing, Winter, ALL,

No SIM does not equal no radio.

There are two basic ways to look at that…

The first is,

“Your mobile phone still fully functions with no SIM in it.”

A point that is easily proved because,

1.1, You can make emergency calls with no SIM.
1.2, The network sees the equivalent of your phones unique network identifier that is the equivalent of an Ethernet MAC address.

Which means,

“Your phone can always be traced on the mobile phone network, even without the SIM in it… And by the faux logical extension the US “Intelligence Community” have been known to use with drone strikes so must you.

Likewise prosecutors have falsely claimed the geo-location of a person by where the phone company says the phone is, which can be hundreds of meters from where it realy is[1][2].

But what would be the second meaning we can take from the statment.

Secondly though apparently ubiquitous radio systems do not require mobile stations to have an ID built into them to work. That is most analog walkie-talkies and mobile transceivers transmit “anonymously”. So for that matter do point to point links you can set up using microwave systems IR laser or similar diodes.

But there is a failing with mobile phones and their suppodadly unique ID’s, which is “it has to be programed in”, and these days[3],

“That which can be programed one way, can be programed another”.

I know it’s obvious to say but so many people forget this or deliberatly ignore it (prosecutors and the like). You can after all buy “hats” for Arduino’s and Raspberry Pi’s that enable you to make your own “4G Phone”, that you can also clear/change such numbers and send bogus GPS readings with… Likewise you can get GPS modules or just the chips, which means if you know where to look, there are some intetesting quite low cost devices comming out of various places not just Asia.

Does anyone remmber the “Talking Rock” in a Moscow Park[4] from a decade and a half ago? Well ten technology generations later you could put similar in a stick, or make something altogether more sophisticated. What caused the Moscow Park Rock to fail was not as such the technology, but the poor OpSec involved. More importantly it gave many other people ideas for products.

It is thus more than possible for low power radio systems to “hide out” in the noise of other systems.

Look at it this way for an Internet analog, HTTPS traffic is ubiquitous, so much so that it’s signal is now the noise. What makes each HTTPS packet different is the IP Header. If you can remove this information or make it worthless[5] then your special signal is just one of many on a network segment as incomprehensible as the next. It’s a trick Tor tries to use as do other Mix-Nets all with varying success but there are many other possabilities[6].

But anyone with a scanner can tell you that sending data over a radio walkie-talkie is realy quite common now there are various Digital Modes used for “Trunking” which enables multiple users to use the same radio frequency but apparently independently of each other. One well known commercial systrm is called “DMR” and it alows for “encryption” thus your ability to tell what is sent down the encrypted channel is difficult[7]. Especially if like the use of HTTPS on the Internet you put your own much more securely encrypted data down it that looks like voice data. Encrypted DMR is now very prevelant on many walkie-talkies you can buy for around 150-200USD and many many “Private Mobile Radio”(PMR) users use not just DMR but use it with encryption turned on because unlike “SelCal” they do get a degree of privacy, in what in some places are very very crowded and overloaded thus “noise rich” environments.

So what is to stop you piggy-backing onto such systems with a covert channel of your own?

I could go on, but the point is, even under extream universal surveillance those doing the surveillance are “resource bound” and that gives all sorts of wriggle room if you know how to set up covert channels.

Such covert channels if implemented with care will remain unnoticed, untill “Poor OpSec” or “loose lips” reveal that a covert channel is in use.

[1] Without going into all the background radio-location is not exactly very reliable. Which is why GPS is a requirment by the USGov. However as many tourists observe what their phone sees as the Greenwich meridian and what the strip of metal at Greenwich and it’s acompanying laser show are not the same by quite a distance. But as I’ve demonstrated in the past GPS is not very difficult to spoof so can be “offset” and with low satellite visability in city streets etc it has quite a high error margin as well as not being able to function. Also it’s fairly simple to design a “GPS jammer in a matchbox” –or if you know where to look, buy– that has a range of only a meter or so. If carried in the pocket with the mobile phone the GPS will cease to function but the mobile will otherwise function[2].

[2] There is a quirk that is unavoidable in mobile phones talking to base stations and that is the difference between turning a phone off and a phone loosing signal when for some reason it can nolonger communicate with a base station. When you turn a phone off it tells the network before it shutsdown so the network knows you have turned it off. If however the phone can nolonger communicate with a base the network assumes it is still on in the same general location untill either the phone fails to “re-register” or a primary service (call not SMS) needs to be made with the phone. This difference is something people need to remember when using those RF blocking bags/boxes. That is put the phone in the RF bag and then turn it off.

[3] If you know what make and model to purchase, you can also find out how to do not a so called “Factory Reset” which is nothing of the sort but a “Factory re-work” which is what is used on the production line to realy clear the phone back to a “blank state” etc and clear/change the serial number and ID number etc.

[4] The fake rock the FSB showed was extreamly primitive compared to what can be done now “ten tecunology generations” later. Or that can now be fairly easily done as a “School Project” for an “Electronic Notice Board”. The fun is now the likes of Apple and Google have done their “COVID-Contact” stuff to their OS’s all sorts of fun that was not previously possible becomes so.

https://www.bbc.com/news/world-europe-16614209

[5] There are many many ways to make the so called “Meta-Data” of network packets become worthless. Anyone who has studied the protocols and the state machines behind them can do interesting things[6].

[6] For example to turn the tails on GCHQ and the NSA, they are known to exploit the fact that the “first packet is accepted, but subsequent assumed ‘duplicate’ packets are dropped”. That is part of the original pre IP design thought up by Gordon Welchman to make the networks not just “reliable” but “fault tolerant” even in the face of “enemy action” (watered down a bit after Gulf War I). That is if you request data from a web site it might get duplicated and sent on different routes. So as we know GCHQ and the NSA, were delaying the legitimate packets by routing them the long way around, and inserting theirs so it arives first. Ever wonder what happens to those second packets? Well the important thing to remember is they get delivered to the client, but usually with TCP the network stack in the client drops them not pass them up to applications (UDP gets passed). Now think what you could do if your network stack does not drop an assumed duplicate? There are quite a few network analysers and the like that do not drop them, but make them available for examination… Such network-stacks are available for Open Source OS’s and are actually simpler than the packet dropping equivalents (so commenting out code is an option). Thus you would be able to examine any such “duplicate” second packets. Remember normally even if they were not duplicates, they would still be droped, it’s why the GCHQ and NSA attack worked. But consider what you could do if you could reject the first packet and instead use the second packet… In effect you hide a second communications channel in the first.

[7] https://www.scannerschool.com/what-can-you-do-when-they-encrypt/