Schneier on Security

Clive Robinson • April 30, 2023 6:47 AM

@ Nick Levenson, ALL,

Re : Security is hard, very hard as distance “goes up”.

“A satellite’s security was cracked from Earth in a test; the satellite was launched in 2019, so its security was presumably nearly up-to-date”

It was not “cracked” and it was in no way secure in the way we talk about on this site in the first place.

Like most satellites, the “European Space Agency”(ESA) OPS-SAT was designed for “availability”. As a “cubesat” it also has significant power limitations, rather more so than an Apple SmartWatch of the same period.

Whilst for obvious reasons I can not go into specifics, the security on most “Space bus frames” is,

“By obscurity not cryptography”.

The reasons are two fold,

1, Cryptography in space is a very bad idea.
2, You can not send a repair technician to fix a space vehicle.

Formally “availability” is based on “Mean Time To Fail”(MTTF) and “Mean Time To Repair”(MTTR). MTTF has a significant inverse relationship with “complexity” in the normal sense[1]. And as already noted with space platforms you can not send a technician up to fix it thus MTTR at it’s basest level is longer than the expected operational life of the platform… So “availability” is a hard problem.

With the issues of a very limited power availability and storage, both of which degrade fairly quickly nobody wants to waste CPU cycles on cryptography. Nor do they want the large silicon real-estate cryptography demands that very very adversely effects “reliability” and can not be “routed around” as that would be an inbuilt security vulnerability…

But consider another aspect. To be secure you are looking at needing “100% correct copy” over 128-16384bits for secure pre-QC crypto and potentially several megabytes for post-QC crypto. This will mean a significant change on the communications “link budget” part of which will be “Power and size” of the platform both of which grow exponentially with getting that “100% copy” over any given bit length message.

The obvious solution is to reduce the “block size” down as much as possible. But unless the unicity issue is addressed a reduced block size will be totally insecure. Worse the requirments for synchronization go up.

So security is an issue for space craft that appears effectively insoluble using currently available technology for several reaaons.

As I’ve noted before I’ve a prototype CubeSat sitting on my desk/bench. Part of the consideration involved is how to make things more secure and it’s far from easy.

For instance one of the many issues involved that you might want to think on carefully before you spend a few million USD putting a platform into space, especially a smaller system like a CubeSat,

All security is based on a “Root of Trust” that needs to be kept securely. In Space there is a lot of hard radiation thus keeping a “Root of Trust” stored “securely against loss” is a very real issue (think bio-metrics and loosing a finger in an accident as a poor aproximation).

How do you recover from this problem, without introducing a significant vulnerability?

[1] Complexity can be used in two ways, the most common or “normal sense” is to increase function or potential function. However as most know a rope whilst it can hold a heavy load requires carefull attachment to the load via various loops and knots. However if the same rope is more more comolexly looped and knoted in a particular way, you end up with a cargo net, which significantly ups the ropes utility as well as making it more reliable in use and longevity. Thus it can be seen that the second use of complexity is to increase reliability. However as with a net -v- rope, such increased complexity has increased costs, not just in it’s manufacture but in it’s operational costs and ongoing maintainablity.

Clive Robinson • April 30, 2023 5:25 PM

@ Nick Levinson,

There is a fair bit of difference between a 1970’s design and launched satellite ~1.4m at the widest in geo-stationary orbit and weighing in at a third of a ton and a CubeSat you could easily pick up and hold.

So “apples with apples” comparisons are important especially as the industry is going for ever increasing numbers of smaller lower earth orbit satellites. It’s not just Hellon Rusk creating problems, a falling out between the US and India over Space utilisation caused the US a very rude awakening. When India’s demonstrarion of a working “Anti-Satellite”(ASAT) Missile in August last year under it’s “Mission Shakti” stuck a quite,large “Keep of our grass” sign in the ground and the message behind it is still reverberating.

But… One of the things not being talked about much outside of closed doors is “Active Denial Of Service”(ADOS) attacks against satellites, which has come to the fore since the incidents to the East of Europe.

The problem originally was the various “Autonomous Decommissioning Control”(ADC) systems that are desired to be put,in new satellites as a pre-condition of launch, to augment the “Post Mission Disposal”(PMD) satellites should undergo within 25 years. The problems are two fold, firstly is such requirments are voluntary not mandatory[1] hence the pre-condition for launch, and secondly the desire for them to be as fully autononous at the satellite as possible, and especially if command and control is lost…

Well the world has changed a lot in the aerospace and satellite industries in the last year or two and it’s not just Hellon Rusk’s StarLink creating news headlines…

As you might be aware from not just Astronmers being up in arms over StarLink’s light polution, others are much more obviously talking about the issue of “space junk” with Hellon Rusk’s failure to orbit and in orbit issues. Stopping the build up of space junk is becoming a major concern along with ongoing research into de-orbiting either into the Earth’s atmosphere, or in the case of geo-stationary sats moving them upwards into “parking orbits” (which at best is only a short term solution).

Less well known is sats can go catastrophicaly wrong and do very strange things and sometimes this causes them to break up into a lot of bits (NOAA-16 for instance in Nov 2015 broke into around 200 parts that we know of from a presumed on board explosion, with 136 large enough to be tracked thus “a significant hazard”). These bits then spread out over a larger and larger volume due to simple orbital mechanics. Eventually they will hit other space platforms causing a cascade or chain reaction that could “close space” to mankinds use.

So over a period of years increasing disposal requirments have been added to a set of recomendations[1]. Which is where ADC systems have been proposed with intent to make them a pre-condition to launch by the UN.

The nasty issue that is coming up is the duality of “autonomous action”. The crux part of which is loss of “stay alive signals” triggers ireversable self deorbiting…

After a moments thought you will realise that simply jamming the stay alive signals is all that would be required to have a 1/4billon USD investment to be irretrievably lost.

Further what is not publicised is that such jamming equipment can be built by an individual at home on a relatively modest budjet. Worse such systems already effectively exist for the likes of “Earth-Moon-Earth”(EME) communications. Also known as “moon-bounce” Amateur operators have very high gain antenna systems to focus RF energy at the Moon to use it as a reflector which enables the signal to be received over the entire Moon facing side of the Earth and similarly most satellite orbits as well.

The only thing stopping such systems being used for ADOS is the supposed RF power limit Amateurs are supposed to obay (voluntarily as enforcment by the FCC etc is effectively nill). However in some parts of the world 1.5kW amplifiers are common and making 5kW amplifiers with LDMOS devices whilst not exactly trivial is fairly easily done (you can already easily find adverts for 5kW “boots” for “truckers” CB…). But… With a little care you can also build “phasing harnesses” that will combine the outputs of several amplifiers in various ways. If the antennas are driven in independent ways then it’s not that difficult to build a “Phased Array” antenna system that can provide a very much more sharply focused beam.

Such things are suddenly becoming talking points in the Space Industry after the recent attacks by Russia on Ukrainian space communications. After all, you can not realistically build an ASAT in your garage and launch it from your garden, but you can make an effective jamming system. Which means any small nation can get in the “Space-Wars” game for less money than they pay individual civil servants and military officers…

[1] The “United Nations Committee on the Peaceful Uses of Outer Space”(UNCOPUOS) along with other agencies such as “Inter-Agency Space Debris Coordination Committee”(IADC) input came up in 2002 with the “Space Debris Mitigation Guidelines” that were further updated in 2007. Which gives us the “25-year rule” which requires satellites undergo “Post Mission Disposal”(PMD) within 25 years maximum, with the prefrence of “immediate removal from Earth orbit”, which has already caused issue in US Courts.

Unfortunately these guidelines are voluntary and only reflect the aims and objectives by several national and international space administration organisations. What some regard as “Over zelous” behaviour by the likes of the US, has led them to ignore other voluntary guidelines, and as more launch capable nations and entities proliferate such voluntary rules are going to be increasingly flouted in part for short term gain as PMD adds significant weight and cost requirments to any mission.

Clive Robinson • May 1, 2023 12:05 AM

@

“Satellite security may not matter much for some satellites, e.g., CubeSats, but some other satellites are valuable targets and so security will matter for those; and then resources will go into better security for them to keep them valuable, as has already happened.”

Satellite security matters for all satellites that have function that requires any level of command and control. Thinking otherwise is a very dangerous tging to do.

The cost of even low earth orbit is over 10,000USD/kg and rising. However the size of payload and max orbital hight is increasing, but ~30,000,000 is what you would be looking for after the cost of the payload to get it up. It’s why some people are looking at systems to launch 5,000 or more CubeSats in one launch with potentially 10,000 CubeSats being possible.

As I said I’ve a prototype siting in my “work area” that could after moderate testing be launch ready. It’s current payload line up is a “Low Probability of Intercept”(LPI) multichannel transponder as well as StarFix optical system. The same payload less the StarFix is also good for attachment to low cost drone or balloon systems.

The smallest usefull Sat I’ve designed as a demonstrator is not much bigger than a large diameter rechargable coin cell. It has been tested to the point we know it can be launched from a flare launcher and drift down over a period of about half a minute, certainly long enough for it’s high bandwidth transponder to get quite a volume of data in both directions.

So the size of satellites is realy not important to function, but is to power systems and longevity.

Which brings us to,

“some low-tech low-cost attacks disable some high-tech high-priced targets; but that tends not to last long if the targets’ owners care about it and put resources into it, which is likelier for high-value targets.”

But what do they do when the rules prevent them from puting in resources?

As I noted earlier the “autonomous deorbit” rules are based on certain assumptions, like the length of time loss of “keep alive” or just “no valid commands” will cause the autonomus action to “kick in”.

Back in the 1980’s every one was talking about “brown out” or “watchdog” circuits.

Essentially they were a single IC like a CMOS 4060 wired up as a self oscillating resetable timer. If you did not reset it it kicked the CPU reset pin, so supposadly causing a restart of the code… That was the idea but it was quickly found to have a whole lot of problems, so became almost obsolete by the turn of the century, with some Microcontrolers having way more reliable Watchdog timer circuits built in. The primary issue that people bumped into was that to often they did not test their software correctly so the watchdog would either not get reset, or not reset correctly, thus fail to work as expected.

One big aspect of not working properly was with communications. We talk about asynchronous and asynchronous communications, but as a rule of thumb that only matters at the bit or byte level. At the high level all communications is asynchronous. That is you can not guarantee getting a data communications packet in any time frame.

This does not mean that the receiver in the satellite is at fault, but loss of communications generally does imply something is at fault somewhere. From the perspective of those making the rules this is a “loss of command and control” thus the satellite should be autonomously deorbited…

The question thus boils down to when should the autonomous system kick in?

From an operators perspective “never” as it’s a high value asset, from the space agencies perspective as shortly after an assumed fault as possible, with some talking 12hour or less.

A big problem is Space Weather can cause knock-outs that can last atleast a week, maybe more we just don’t know. Further terestrial weather can cause ground stations to be out of action routienly for a week or so. Unexpected hurricanes, lightning storms etc can cause grid outages for two or more weeks are not unknown with a couple of months not unknown (ask people in Auckland NZ, back in 98, the city was out for five weeks, and there have been similar since with Henderson). There is some evidence that some types of space weather can cause tetestrial weather to follow after a few days. But likewise you get one failure throws excessive load onto other circuits/systems and fault after fault happens, sometimes as a chain of failures other times as a growing cascade that increases rapidly (ask customers of PEG in California about their deliberate shut downs when the wind blows just a little). In a laughable series of events a solar observatory used for making predictions for space weather predictions for US grids, got shut down by a power cut for quite some period of time, because the prediction data became unavailable…

So LOS could be a month or more… But from an agencies perspective how much damage could be done in as little as a couple of hours without a Command and Control channel?

With Crypto based security the chances of LOS goes up significantly, irrespective of the size of the satellite, or length of time it can operate on batteries with only partial charging and the like.

The realy bad habit of people in managment waving issues away with the false optimism of “It will be alright on the night” or equivalent, is not sonething responsible engineers will willingly tolerate. But managment cut their pay checks and hire and fire in ludicrous ways and politicians just take a back hander from managment to look the other way… Which is why we have so many “Acts of God” that cost so much.

Clive Robinson • May 1, 2023 11:40 AM

@ Dusapin,

SOHO was a quater of a century ago.

The thing it proved more than anything else was the redundacy of unnecessary hardware.

That is it was built with gyroscopes it did not need.

But it waa a large system and it’s command channels were not exactly secure even by 1980’s standards.

But I’m not sure you actually understand the issues of lost “Root of Trust” and built in vulnerabilities when you say,

“The crypto subsystems usually have an hierarchy of keys, with built-in fallbacks for some and upload procedures for others.”

…

Clive Robinson • May 2, 2023 3:33 AM

@ Markc,

Re : Cipher Feedback mode.

“You might consider encryption using CFB mode. It’s light weight and self-correcting especially with single bit feedback, it can even recover from phase errors (bit shift).”

It is something that I’ve looked at in the past “as part of the solution”. That is it can be used like a “matched filter” sychronizer in Spread Spectrum and other “Low Probability of Intercept”(LPI) systems.

However it has two problems,

1, It does not protect against tampering.
2, It still requires a root of trust.

The first indicates why it’s only “as part of the solution”[1] as it requires further processing via some form of detecting / “correcting codes”.

And the requirment for a fragile “root of trust” still persists.

The solution I suspect will have to involve the use of matched filters and error correcting codes and the equivalent of ‘building a reputation”.

[1] From Wikipedia,

“The block cipher modes ECB, CBC, OFB, CFB, CTR, and XTS provide confidentiality, but they do not protect against accidental modification or malicious tampering. Modification or tampering can be detected with a separate message authentication code such as CBC-MAC, or a digital signature.”

Clive Robinson • May 3, 2023 10:34 AM

@ iAPX, JonKnowsNothing, Winter, ALL,

Re : Tor is not what it claims, not even close.

“I criticized a subject about security, with an “expert” dubbing Tor as secure.”

Anyone who thinks Tor is giving them security of any real use is “wrong”, sometimes “dead wrong”.

Our host @Bruce once used an analogy of a young girls diary with a lockable clasp on it. It might keep her little brother out[1] but nobody else. Well that’s kind of a reasonable description of Tor.

“Tor might keep the script kiddy next door out, but if even the FBI can succeed at exploiting it’s weaknesses and they have, how secure is it against say a SigInt agency?”

Over the years I’ve listed many failings of Tor on this site and got a lot of kick back from Fan Buois who probably new more than your,

“young security ‘expert'”

Who was apparently anything but…

I’ve even described how to fix some of Tor’s many failings, and still got kick back…

Over the years the failings became more widely known and the kickback less.

But one thing that will never ever go away with Tor in the current way it works is you are painting a big red target on your back by using it.

So as @JonKnowsNothing indicates,

“TOR gets promoted because of the perception is is Secure and Safe. It is not.”

It’s traffic is obvious and almost all of it will have been hoovered up by “collect it all” and put in storage untill they need to break it in any of a number of interesting ways that probably do not need the cryptography side of things to be broken.

Which brings us to @Winters point,

“Tor helps in some cases, not so in others. Where it helps, there are few alternatives for non experts. Where it does not help, it is for people that already know they thread on dangerous grounds.”

What @Winter does not mention is the huge asymmetry between “helps in some cases, not so in others”. It’s actuall level of “helps” is just a tiny tiny fraction compared to the “not so”.

Even after a decade of knowing what needs to be fixed and they realy are not difficult fixes, they’ve largely not happened so a valid question to ask the developers is,

“Why not?”

The answer to which appears to be something along the lines off,

“The users want low latency for their naughty video watching.”

And that just about covers it…

[1] Sadly not all little brothers… Even though my big sister did not keep a diary she did have a lockable jewel box. As I’ve mentioned I knew how to pick those sorts of locks by the time I was eight… So moving her jewlery around in the box was a way of reminding her I was not happy with her behaviours towards me. Our parents did not believe her stories untill a couple of years later when they actually saw me pick a car door lock after the keys got locked inside by mistake in the middle of nowhere, lets just say it was a “stony silence” day proving to a young lad that “No good deed gos unpunished”… The reason they did not suspect before that is all I was interested in was picking the lock and then locking it again, so most of the time I never touched anything behind the door, let alone move it about etc.

Clive Robinson • May 4, 2023 4:49 PM

@ &ers,

A bit of history for you…

Peter Sommer alias “Hugo Cornwall” wrote only the first three editions of his book “The Hackers Handbook” the fourth edition was written by Steve Gold (of BT Prestel Hack fame and cause of 1990 UK Computer Missuse Act).

Not all the content in the 4th edition was from Steve’s experience, some was from other peoples “practical experimentation” and reporting. Some like Pete and Steve were considered more than a bit paranoid… It came in some cases of having had the Met Police trying to turn them into criminals or directly knowing others who had (as we much later found out it was on the direct command of UK Prime Minister “Mad Maggie” Thatcher who was selling BT off to fill the “war-chest” to “give away” to certain favoured interests and wanted nothing in the press etc that might reduce the haul).

It’s a funny old world but by 1993 I was very definately “A Proffessional Design Engineer” keeping my nose fairly clean developing communications and embedded systems… Some of which was “Intrinsically Safe” stuff for the Off-Shore industry, some for space payloads, other stuff included “Electronic locks and environment” systems for amoungst others the leisure and entertainments industries… Which ment I bumped into Mossad pretending to be an IBM sub-company (and they realy did not like “electronic locks” being used in hotels…). Then there was the “surveillance electronics” and doing a little “Contracting” on the side which was quite fun. Oh and high power transmitters and some “Crypto stuff” some of which was to protect software downloads to very expensive audio equipment for the broadcast industry…

You could say I had “itchy feet” as I did not settle into a “company-man role” as well as having “One eye over my shoulder” much of the time.

Life’s been interesting on and off including doing research and working at a Uni, that became embroiled in a UK Spy case and had MI5 and the Met Police stomping around in big boots making a nuisance of themselves. But hey as the old joke has it,

“When their ain’t much on telly what else are you supposed to do for entertainment…”

Clive Robinson • May 5, 2023 12:44 AM

@ JonKnowsNothing, SpaceLifeForm,

Re : Atlanta Epidemic Intelligence Service…

What can you say…

I won’t say it’s got the same cachet as the “Hounslow Hilton” with all that luxury feeling of an underground car impound. But when several popular cartoons take the p155 our of it you should take a hint…

Some might ask,

“Atlanta Intelligence, Oxymoron?”

But seriously with a “World of Coca-Cola” in twenty acres given as a top tourist attraction[1]… And of the top 100 worst crime cities in all of North America, Atlanta ties in 12th place with Houstan Texas…

You kind of get a feeling it’s not going to be “fun, fun, fun”…

So you have a quick looksie on the web and what’s almost the first thing that pops up,

https://edition.cnn.com/videos/us/2023/05/03/atlanta-shooting-timeline-police-chief-sot-vpx.cnn

Yup, you have to ask who in their right mind would want to go to Atlanta…

There used to be an expression used when people asked not very bright travel questions which was,

“Hey you wana catch a cold?”

Well I guess it got updated.

[1] Maybe it’s me, but realy does this fill you with glee,

‘https://www.worldofcoca-cola.com

Oh and as for that secret recipe in the vault… We know one ingredient is “gum arabic” from current news on Sudan and it’s hostilities,

‘https://www.wsj.com/articles/key-ingredient-in-coca-cola-chocolate-and-red-wine-is-caught-up-in-sudan-crisis-cf769a87

How long before some ask “Is gum Aribic the new blood diamonds?”

Clive Robinson • May 5, 2023 6:24 AM

@ Winter, JonKnowsNothing, ALL,

The third paper you link via the Science Direct site link of,

‘https://www.sciencedirect.com/science/article/pii/S0167739X23000651

The paper is apparently not due to be published ubtill “August 2023” and the links on the site you give are not functioning for me, so I’ve not read it (have you?).

But from what is available I gather it is not about “finding users” but “Onion Servers” which unlike users are considered by some as part of the Tor “onion network”.

The second paper you give,

‘https://ieeexplore.ieee.org/abstract/document/9471821

What they call “De-anonynising the user”, is not about finding who is connecting to the network from their PC etc but,

“De-anonymising the user is usually conducted with one of two objectives: finding out who is visiting a particular website or finding out what websites are being visited by a targeted user.”

That is tracing backwards the connection inside the Tor “Onion network” from the server to an unknown user. Or by recognising the traffic signiture of the type of information being accessed (that is tracing the signiture backwards from server to user).

But not between the users PC etc and the entry guard a user has connected to.

As I’ve indicated repeatedly, the authorities only need to know you are “connecting” to hang you, and do not need the actual information sent or received or to and from where.

As for the first paper you give,

‘https://www.researchgate.net/publication/341319768_Traffic_Analysis_Attacks_on_Tor_A_Survey

“In low latency anonymity networks such as Tor, an adversary is generally aiming to confirm the source and destination of communication. The current design of Tor’s network assumes the absence of a global adversary that is able to monitor both ends of the communication, entry, and exit guards, and does not provide anonymity against this type of adversaries. Instead, Tor’s threat model assumes an adversary that can observe only a fraction of the communication, and is able to control only a fraction of Tor nodes, either by runningh is own ORs or compromising an already running ORs. Based on the proposed threat model, attacks can be categorized according to how practical their model is, in terms of the assumptions made and the required resources to enabling the attack.”

Is specifically excluding the PC to Guard monitoring by anyone who has access to an Internet “upstream router” etc such as an Internet Service provider between the user PC/device and the entry to the Tor Network “Guard node”.

So again I say,

“Say’s who?”

Oh and please when you respond, don’t just do a quick citation search and give a link based on the title. First actually read the paper and understand what it is they are saying with respect to what I am saying.

As “De-anonynising” to the paper authors may not mean the same as to an agency intent on dragging you in on only minor evidence effectively torturing you debying you legal representation and then “show trialing you” or worse, which appears common even in the US (see previous behaviours of Chicago Police US MSM kept hidden deliberately).

Clive Robinson • May 5, 2023 9:43 AM

@ Winter, JonKnowsNothing, ALL,

Re : “Global adversary” is kid up the street.

“Unless you are up to a Global Adversary, Tor is helpful.”

Realy?

No, as I’ve said your ISP has knowledge of your connection to Tor. In the US they are legally “encoraged” not just to store it but report it to any “official” agency that thinks it’s official. Oh and they are alowed if not encoraged to sell the information to data brokers, who will sell such information for a few bucks (remember the Priest and the Gay Get Together scandle).

And worse in quite a few cases any half smart script kiddy on the same IP address range sub segment as you as well…

Not sure how these people are “Global Adversaries” just numpties on the make…

So again the question comes back to,

“Say’s Who?”

Clive Robinson • May 6, 2023 1:00 PM

@ Winter,

“But the subject is “torturing USA citizens on US soil” for using Tor or other secure communication means.”

Nope, and you know it’s not as can easily be seen by your previous comments.

I stated quite correctly that using Tor was painting a target on your back. That is your connection is easily seen.

You decided otherwise, when challenged you did a ChatGPT syle Halucination and produced three academic papers that as far as we can see in no way contest my statement.

In fact two of the papers as quoted change their meanings of words and limit the scope by saying that certain types of attacker were ruled out of consideration.

Various facts supporting the connect to Tor issue you’ve failed in any real way to addresss.

If you want to use Tor youself that is your choice but to try and convince others using Tor is a safe thing to do is most certainly not, becausr it is not safe, never has been safe, and the way it’s future course is going it never will be safe.

Your “Nothing is 100%” argument is a typical fake-argument used by people who basically have no rational or logical counter argument. Also what is in effect “it’s Tor or nothing” type argument either.

You keep trying to “move the goal poasts” so that you can claim even your “own goals” are winning moves on your part…

My argument still stands and you’ve provided nothing to the,

“Say’s Who?”

Call for you to provide valid evidence to support your claims.

You’ve put yourself on a hook of your own making, now you are trying to wriggle off…

That is your choice of course, but remember people can see it for what it is…

P.S. With regards bike locks, when at school I could and did open them faster than their owners could especially combination locks. I’ve mentioned this before more than a couple of times on this blog. It got me into trouble not because I stole anything, but because people wanted some one to blaim, therefore they would blaim me even if they knew I was not to blaim. You show a security measure to be not secure, then you are breaking somebodies rice bowl and they are going to use you as an example or an excuse to get themselves out of the hole they created for themselves…

So you could say I’ve experience of having had a target painted on my back on rather more than one occasion, so know how some of it works.

Oh by the way comparing your bike and your nations gold, is unless your country has no gold, not comparing apples with apples… is that a LLM type hallucination or conflation?

Clive Robinson • May 7, 2023 10:52 AM

@ Winter, JonKnowsNothing, ALL,

“That was not breaking Tor. That was bad opsec.”

Using Tor “is Bad OpSec”…

You are painting a target on your back by using it.

As I’ve repeatedly said in the past using Tor is a bad idea as you can not hide you are using it. That is just a part of the “Bad OpSec” involved with Tor.

There is also the myth of Tor’s “Mix net by Onion routing”. The low latency and only using an entry guard point and an exit point the network locations are commonly known and a single intermediate point means that there are real issues that need to be considered.

If an adversary can see entry or exit points all the traffic can be mapped thus the intermediate points found. Due to the very bad design of Tor especially the low latency user traffic through that intermediate point can be mapped in both directions.

The fact the “physical network” and the “logical network” of the Internet are far from being the same and there are well known “choke points” on the “physical network” which aid in tracking low latency Tor traffic in the logical network means that “Traffic Analysis” can say a lot about the message content, without needing to see the user message content.

I could go on but it’s becoming increasingly clear you’ve painted yourself into a corner based on incorect assumptions… To the point you are now in effect handing out dangerous information as others not just me are gently trying to tell you.

Clive Robinson • May 7, 2023 6:07 PM

@ Winter,

Re : Binary is not the only selection.

“A young woman searches the web for abortion in some US state. Should she better use Tor or not?”

That is a very silly argument to make as it shows either your lack of understanding or a deliberate presentation of a false argument.

The choice is not “Tor or Nothing” and never has been as I’ve clearly stated in the past there are way way beter ways than Tor with it’s multiple failings.

Which brings us to your assertion of,

“Which is ridiculous when referring to your example which had no overlap at all with Tor.”

It’s a fact of life for every one using Tor as is, so you obviously do not understand what you are talking about.

As for,

“All your examples require a Global Actor like the NSA et al..”

I’m seriously begining to doubt your basic comprehension.

I’ve already explained how anyone from an inefective script kiddy upwards can find out if you are using Tor or not. They do not have to be the NSA or what you call a “Global Actor”. Your ISP knows every primary connection your PC etc makes through their supplied service, and they know who you are through billing etc. That applies where ever you live and have a home ISP you use, it’s not something you can argue against it’s a simple and uncontestable fact based on the way the base protocols of the Internet work.

Further in many countries not just in the US or UK but many other Western/First world nations, there are two things that happen,

1, The Nation State requires such records be kept.
2, There is not a prohibition on such “business records” being sold.

Thus for pocket change your primary connection to a Tor gateway/guard is available to just about anyone.

The fact you are not taking this on board would be rather worrying in of it’s self. But the fact you try to then misrepresent that using Tor in the way most people do is not “Bad OpSec” actually makes it highly alarming.

But I guess you are just going to come back with more factually incorrect statments, or deliberate ignoring of why Tor makes people much more vulnerable than they need to be…

Your choice, but so far you’ve been wrong in what you say, and I somehow suspect if you carry on, then you will carry on being wrong…

Clive Robinson • May 12, 2023 3:29 AM

@ MarkH,

Re : Orbits and station keeping.

Now this thread has quitened down.

“I suspect that many satellites that “move” (geostationary is a different case!) can be adequately controlled once per orbit.”

The general idea of “station keeping” is to “plan ahead”.

As much of orbital mechanics is predictable over quite a period of time with simple geometry you can make a map and stop watch solution. The time of orbit gives you the satelite “effectctive hight” and what the current sub-satelite point is.

Thus as an orbit decays the satellite orbit time gives an indicator of when the satellite will need to be kicked-up into a higher orbit, by how much and at the most efficient time (often the Hohmann transfer uses the least amount of propellant in moving between orbits, however in some cases bi-elliptic/patabolic transfers are better but at more than double the transfer time…).

However orbit decay in lower orbits is mainly due to friction with the atmosphere. The Earth’s atmospheric hight in the upper regions is due to “Solar Weather” which is a direct result of the output from the Sun.

Solar Weather has a basic 11year cycle, and the Sun output tends to be related to “sun spot activity”. Such activity though appears quite random in the short term and is thus effectively unpredictable. Which can lead to very short decision times on station keeping.

More complicated is the effects of other orbiting bodies… As is fairly well known Kepler got the two body problem fairly well nailed down. However nobody has got the three or more body problems to “play nice”, and our solar system is full of orbiting bodies, that occasionally get stired up by “visitors dropping by” some of which trundle through the space defined not just the Moon’s orbit, but inside Geo-Stationary satellite orbits, with some burning up in the atmosphere or making holes in the Earth’s surface.

Thus the overall effects can be unknown even in very short periods of time, and requiring almost “play it by ear” corrections.

With regards,

“Obviously, if a shared secret is lost, then control of the satellite can be lost also. What can be done to prevent that?”

Well the Chinese have solved part of the problem…

Clive Robinson • May 12, 2023 7:13 AM

@ MarkH,

“With the capabilities of present technology, maintaining integrity of key material over a satellite’s lifespan appears to be a matter of careful engineering design.”

As I’ve previously noted,

“maintaining integrity of key material”

Is not sufficient especially with Post-QC sized keys and current “staleness” requirments. As I mebtioned in,

‘https://www.schneier.com/blog/archives/2023/04/friday-squid-blogging-more-squid-camouflage-research.html/#comment-421340

As a reply to your Mayday post of 14:09.

But you make a quite invalid assumption of,

“Who would be surprised that a satellite costing more than 1 billion USD incorporates more safeguards than one costing around $100,000?”

Price has little or nothing to do with it, but time does. There are very few 1 billion USD satelites and rhere design / manufacture cycle is long so their specifications are likely to be much older than that of a cube sat.

But in either case the “storage of KeyMat” for 25years is not the solution to the problem.

As I’ve said the Chinese have already solved it in a way that does not require storage of massive quantaties of KeyMat as became evident with the recent “balloon shoot down” incidents.