Schneier on Security
A blog covering security and security technology.
« Self-Propelled Semi-Submersibles |
| Billboards that Watch you Back »
February 11, 2009
Cloning RFID Passports
It's easy to clone RFID passports. (To make it clear, the attacker didn't actually create fake passports; he just stole the data off the RFID chips.) Not that this hasn't been done before.
I've long been opposed to RFID chips in passports, and have written op eds about them in the International Herald Tribune and several other papers.
EDITED TO ADD (2/11): I got some details wrong. Chris Paget, the researcher, is cloning Western Hemisphere Travel Initiative (WHTI) compliant documents such as the passport card and Electronic Drivers License (EDL), and not the passport itself. Here is the link to Paget's talk at ShmooCon.
Posted on February 11, 2009 at 5:09 AM
• 62 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
And here is a nice app for making a backup of your (anyone's?) passport data using your Nokia NFC mobile phone: http://www.dexlab.nl
I certainly agree that the cloning must be easy and I don't see the value in the chip, since it would be irresponsible not to confirm the information with visual inspection, but...
So what if you clone it. Is it similarly easy to clone the rest of the passport? It's actually very hard to do that.
Ah. A new use for cigarette cases (remember them?):
Keep all RFID-equipped documents inside, thus making them resistant to cloning...
One more thing to trip the magnetometers!
On a related note, I needed to be wanded down because I'd forgotten a pop-pack of Sucrets in my sportcoat pocket. Not at the airport, but at a court-house.
Lucky thing I wasn't arrested before I could plead my case regarding the traffic ticket...
A researcher at Shmoocon last weekend said that they tested some of the products being sold as protecting RFID docs...all lame. Tin foil works better at attenuating the signal. He also said there were a ton of variables but they were able to get a reliable range of 10 feet and a max of 100. You can copy a bunch of data in 10 feet on any city street.
Then we had a nice discussion on the banks incentives. They don't eat the loss the vendor does. They collect transaction fees on sales whether they are legit or fraudulent. This should probably change...
OH! And, Arguably, with the Pay Pass system the Credit Card industry is violating it's own PCI standards.
I was thinking a wallet with a copper foil lining. But that would probably get attention in an xray and a metal detector. hmmmm if it ain't legal to take police's pictures (in England) or legal to encrypt your data comin' through customs...logically...a countermeasure to protect oneself from the cloners will provoke a resolution to make it illegal 'cause it interferes with security measures. House Resolution Nr ??? anyone.
Are we fighting the clone wars?
You know I disagreed with the court decision to lift the gun ban in DC but now I see the merit in the argument we have an inalienable right to protect our selfs. And if my data equivilants aren't part of me I don't know what is.
This articles is about cloning the new US Passport (or PASS) card, which enables travel to Canada/Bermuda/Mexico. Not quite the same thing as a passport...and containing a different type of RFID device altogether.
If they just stick the passport on the reader without opening it and then just wave you through "cus the computer is never wrong", you don't need to clone the rest of the passport....
Anyway I had one of the first RFID passports. I have only gone through 1 airport that could read the chip data. We got held up for 20min while they used the first RFID passport to come through as a training exercise. The next time i came through that airport they didn't use the RFID reader at all....
What is the best way to disable the RFID without visibly damaging the passport?
RFID is seemingly already obsolete. The thought, I suppose, was good, but the usefulness is severely limited in day to day situations.
In my "dayjob", we require people to have, bring and display their passports for inspection, and not once have I ever heard of any department that uses the RFID for any kind of verification. It's all about experience, the human eye (and a magnifying glass) to verify suspect passports.
@Bob S. - you should never disable or alter any part of your passport yourself. That can get you in trouble with all sorts of agencies in a lot of places - including your own immigration authorities, and in a worst case scenario land you in jail (not for a long time, but any time in jail is an inconvenience to most people...)
If the RFID fails on my passport how can that get me into trouble? Or for that matter, how will anyone know if the passport or the reader is at fault?
Bob S was wondering if there was a way to render the RFID inoperative so that cloning is impossible.
@BF Skinner -
"A researcher at Shmoocon last weekend said that they tested some of the products being sold as protecting RFID docs...all lame. Tin foil works better at attenuating the signal."
Do they suggest we'd best keep this under our hat?
*Movie Plot Terrorist Alert*
Instead of cloning passports, read and analyze the data. For instance, place a device which automagically reads the passports of folks around it and detonates your bomb if there are at least 2 Americans in range.
@Larry's "there are far easier ways to do it"
Unfortunately, that's obvious to everyone but federal, state, and local officials that panic at the sight of every unattended backpack and any movie plot "threat" regardless of how ludicrous.
my understanding was that the best way to disable the RFID chip without visible damage (microwaving will leave burn marks) is to use a hammer to smash the chip.
"Is it similarly easy to clone the rest of the passport? It's actually very hard to do that."
Right, which is why the RFID is really stupid. If the RFID checks out, the rest of the passport is less likely to be examined carefully.
@ Bob S, et al,
"What is the best way to disable the RFID without visibly damaging the passport?"
First of, if the RFId is not working your passport is nolonger valid for travel.
Which would be embarising if you had just spent a shed load of money on going away on holiday and instead of enjoying yourself got put on the next plane back to where you came from...
And the reason a microwave leaves burn marks is to much power.
As has been found by some people an HF signal close to the RFID frequency (~13.5MHz) will if it has sufficient power burn out the chip in a not visually aparent fashion.
But as I said it's not valid for travel and in the UK that means coughing up 80GBP (around 120USD) to get a new one...
As I have said repeatedly in the past RFID Passportss actualy make the life of criminals and terrorists easier.
Because it makes life fairly easy to go "ID Shopping".
I'm just waiting for "ID Brokers" to set them selves up and trade RFID Images in the same way credit card info and personal ID info is currently traded...
Sometimes throwing technology at a problem makes it worse not better and in the case of ID RFIDs are definatly a huge step backwards...
@bob "Do they suggest we'd best keep this under our hat?"
The researcher did say that he wasn't advocating that but the crowd just nodded and adujsted their beanies.
Also odd is that even though there's a certificate in the chip in the passport with biometrics?
CBP isn't reading it at entry points.
so it goes
To clear up a couple of things:
- This demonstration was not against the Passport, but rather the Passport Card (as well as the Enhanced Drivers License and other components of the WHTI). There's very different RFID technology in the two documents; however bad the Passport is the WHTI documents are infinitely worse.
- My full presentation is available online and should clear up a few things: http://video.google.com/videoplay?... I would expect that most people who follow Bruce's weblog would find it interesting.
- There's a lot more to this than just cloning RFID tags. Watch the presentation to see why we have an identity card system that can be tracked to a theoretical limit of 2.5 miles, despite the objections of our newly-appointed Secretary of State...
It seems so obvious to me: RFID is just a barcode that doesn't require line-of-sight. The barcode should not contain ANYTHING other than a database key--any biometrics should be stored in a secure database and merely referenced by the RFID, not stored on the RFID.
"any biometrics should be stored in a secure database and merely referenced by the RFID, not stored on the RFID."
It won't work for a number of reasons.
The primary one is that the majority of people who will be reading it (Police / Banks / health care accountants et al). Will be so numerous and dispersed that they cannot all be connected to the database at the same time.
The communications network required for 100% reliable access would make the credit card network pale into insignificance by comparison.
Also the communications from hand held devices by the police and other mobile "checkers" could be easily jammed. Unlike the credit card where it is a lost sale or financial risk when their network is down, an ID network would if down be a significant risk.
Then there is the cost of the readers having to talk to a network this is likley to be significantly more expensive and less reliable than a simple smart card reader microprocessor and LCD display.
Then there is the cost of maintaining the database in a state that even vaguly approximates something secure.
Also big databases of any form are natoriously difficult to maintain.
And how about the question of trust? You might trust your safe keeping of your card but would you trust a "political target" driven organisation on a Govenment contract to not employ the wrong sort of people...
I recently got a new passport with an RFID chip. I casually suggested putting it in a microwave to the public servant at the municipality who gave it to me. He replied that a hammer would be the tool of choice. And, of course, he warned me that I wasn't allowed to do this...
Off topic but...
I know I have the habit of "bleating on" about fragile infrastructure and how you just need to know where to stick a leaver to get a greater force to break things.
But it keeps on happening in this case sending a "." To DNS servers with a faked source address,
Great presentation if you have time to watch it. Since the current cards are writeable if you have the lot code, wouldn't the best way to invalidate the whole program be to re-write everybody's card.
The speaker states that if he determines the lot code for the cards, the card is re-writable. Since the cards are sold in bulk, presumably there will be a (probably very large) range of valid tags. Just brute force the write attempt against every previously seen lot code. The card does not need to be read to rewrite it. Just set up the device to continuously cycle through all of the known lot codes and overwrite the tag with generic information and leave in a secure high volume location (ie 5 miles from the border a shopping mall or McDonald's) and run it during rush hour each day.
This is more valuable than zapping cards as it invalidates the entire system - at least until they get smart enough to use write-once gen 1 cards for what is supposed to be a secure reliable ID. Why would a secure ID card ever be rewritable if you expect to maintain trust in the contents.
I was the very tall man with the mustache.
But I was actually next door at Hackerbot Labs' Johanson's RFID Unplugged session.
Bummer they were both scheduled at the same time. Got the DVD but haven't watched it yet.
Shmoo should embrace PiP next year.
The reason why it will get you in trouble is that if I get your passport in hand, check it over, see that the RFID chip is burned or broken (a hammer assault on it will most definately make a mark that I can see)
I can also see in my fine little x-ray device if it's been physically broken. Anywhere with one RFID reader will have another one, most likely, so checking the reader is easy. For that matter, check a couple of other chips through it and you'll know if it's the passport or the reader...
Why is nobody discussing the issue of protocol, like an authenticated Diffie-Hellman?
I really wonder why the question of appropriate encryption protocols, like authenticated Diffie-Hellman is not being discussed in the context of eavesdropping?
"An appropriate protocol will not allow to eavesdrop anything:"
The real problems with RFID IDs/Pasports/etc start well below the "data security" level.
The simple fact is an RFID at a very minimum indicates it's existance before it even starts receiving let alone transmitting information. This is due to it taking it's power from the RF field.
If you want to know how this is done with a very simple electronic circuit have a look at,
However, if you automate the GDO functionality and take the "dip current" change into an averaging circuit then you can analyse it's waveform.
And in a similar way to "differential power analysis" you can determin other information, such as the RFID chip manufacture and possibly it's stepping. This is similar in concept to netcat enumerating a TCP stack and determing the OS of the system.
However you can also by offsetting the frequency determin the charecteristics of the RFID "tuned circuit" pickup coil. This provides further information about the device the RFID chip is in.
This knowledge then tells you a lot of information about the ID device even before it starts to communicate. For a number of attacks it is "game over" at this point...
Further the first part of the cloning process is carrying out an autherised communication in reading the data so it is not "evesdropping" on the communication path.
So the level of security of the communications protocol is a mute point from these asspects as it would not stop either attack.
With regard to,
"Does this look like a step towards a solution? Opinions?"
About a secure OS for smart cards, you are again looking at the wrong level and mode of operation.
In many ways the "Welsh Castel" OS is a solution looking for a problem to solve in the electronic ID problem domain.
The Cloning of the card data used two autherised actions within the functional specification of the RFID ID Protocol.
A secure OS cannot fix a broken functional specification only changing the specification can do this. In effect it is like the WEP problem again, the engineers and others who designed the RFID ID protocols did not have the relevent problem domain experiance.
@ Clive Robinson
Well, this is not only an operating system for running a protocol, but if you look at the public info about it, it also uses different kind of hardware.
I am very familiar with the grid-dip like methods and the physical level.
The physical level will not indicate presence and internal workings, because the hardware runs off a "constant current generator", and the hardware contains measures against DPA.
The serial number and batch number doesn't help you anything. The card contains a modular arithmetic engine, a true random number generator, and the usual components of a security processor. The card also contains exception sensors for unusual operating conditions, frequency, field strength, temp, injected faults. This is all known.
It is indeed a miniaturized security processor like the IBM 4758.
As such it can run a protocol which is based on authenticated Diffie-Hellman.
The simple cards, which have been attacked, don't have such hardware and don't allow to run such protocols.
This is the next level of the game.
So far it was only kindergarten.
This will also allow protocols to include Visa-Stamps, Border crossing tags, past communication history, updates and biometric info, finger print etc. to be bound together by signature.
It may be possible to clone the info by physical analysis, but you cannot forge a new identity by separating the photograph, fingerprint, name, history from each other to make a frankenpassport.
In particular such hardware is able to generate its own key pair and give out the public part and have it certified together with the other components
The nice part is also, if you clone by physical analysis, the clones will diverge by history, so you have two passports with exactly the same name, number, id, picture etc but they were run through different update authorities, at different times, they have different communication histories, giving a different hash value.
At this time you may notice that something is wrong. And also the chopped off finger in your pocket doesn't look and smell as fresh as it used to be...
You can also make a passport, that requires you to check into some Gestapo-Kiosk daily for an interview, or it turns you in whenever it is checked.
Of course the running number of your daily interview sheet in that kiosk is bound with your other info in the passport by hash, and included. So the border agent asks you what you had for dinner yesterday, checks it in the computer - ok: Pea soup.
So you see: "Your mothers maiden name" is just lame!
So, googling the IBM 4758, I can only guess that attacks to such passports will have to follow the foot steps of attacks on the IBM 4758, just look that up and see what has been done/can be done. This is going to be interesting!
As the processor has generated its own key pair, and has the public key bound and certified together with the other information, now the public information will not allow you any more to clone a passport, as it does not contain the private key, matching the public key, which is certified. So a clone, duplicating the public information, cannot forge messages of the passport, because it cannot sign with the matching private key. In the result you have to resort to physical analysis to get the private key, to make a cloned passport. But Frankenpassports are still impossible, because the information is bound together and cannot be separated, even if you resort to physical analysis.
I hope this summarizes the perspectives correctly.
"IBM 4758, just look that up and see what has been done/can be done. This is going to be interesting!"
That is for some distant far of future where you and I will be to old to care about it...
Why, you can say it's similar to the "MicroSoft Legacy Software Problem".
MicroSoft did a quick and dirty hack to get MS-DOS up and running. Likewise with Windows. Neither had any kind of security etc considered let alone built in. MicroSoft could do this because they where effectivly the only "recognised" game in town.
Third parties software organisations then developed their own programs to run on MS-DOS and Windows and did things the way that suited them not MicroSoft or their customers.
To avoid alianating customers MicroSoft had to effectivly support the legacy third party code with "backwards compatability" through future generations of OS and Windows.
Along with this support came all the security insecurities we are still seeing and suffering a quater of a century later...
RFID Passports are based on an International Standard which apears to have been ill thought out. And pushed through with unsemley haste at the behest of one interested party for political expediancy. This was only possible because there is no equivalent system in place. So, is very much like the MS-DOS and Windows hacks...
From this came functional specifications that in their turn where ill thought out. From these functional specifications came detailed specifictions for the interoperability of products. From all of this was developed a costly infrestructure with component parts with expected life times of ten or more years.
Now it has been foisted on nations and they have had to develop their own solutions. They have become the thrid party suppliers who will not want to change simply because somebody says "we got it wrong".
So now the current RFID system is in place how long before it has sufficient inertia that it cannot stop rolling?
At a guess it will take another five to ten years to identify most of the weaknesses in the current systems. It will take another five to ten years to come up with an agreed revised international specification. About another five years after that the first systems will start to roll out but must maintain compatability with the old systems for atleast another ten years after that. So between a quater and half a century from now the current weak pasport system will still be in use and exploitable...
Personaly I suspect that due to heal dragging by many countries fifty years may be a little optomistic...
And then of course there are the unknown unknowns still to come as well as the unknown knowns. Irrespective of the technicological issues there are the messy human ones.
For starters most people cannot prove they are actually the person the pieces of paper they hold say they are. Then for instance will nationality based on where you are issued a certificate of birth actually have any meaning on identity in fifty years time? Will there actualy be over two hundred and fifty independent countries in fifty years?
Less than a century ago passports where not required for traveling except where there was assumed status such as a diplomat etc. Nationality was fairly irrelavent as well as long as you paid your dues (taxes etc) where you were located.
Just over fifty years ago people were denied nationality or status simply based on their religious or ethnic status and in some parts of the world this still happens.
Less than fifty years ago people where denied passports for ideaological reasons and prevented from traveling and this still happens.
What effects will things like AIDS, Religion, DNA, Corperate globalisation and education/skills have in fifty years time, some are already used to decide on if you can work in / migrate to other countries.
I suspect the whole notion of an electronic passport needs to be re-thought from a very very fundemental level and detached from all current assumptions based on paper passports and their resulting support systems.
A short term view on a current technological solution starts to look a little irelavant when viewed this way.
Well, the "IBM 4758 like functionality in a smart card" is available as a smart card chip, and you can buy it.
In the link I gave it says:
Lastly, the Caernarvon privacy-preserving authentication protocol  protects a smart card holder’s identity and has been incorporated into the European CEN standard for digital signature applications on smart cards . It is currently under revision by CEN  and will be submitted to the International Standards Organization (ISO).
An initial implementation of the Caernarvon operating system is fundamentally complete. It consists of approximately 33,000 executable statements written in C and 14,000 executable statements written in Assembler. The low-level design documentation is incorporated in the source code. In addition to the executable statements, there are approximately twice as many lines of comments. The software has been tested on a hardware emulator. If printed and stacked, the Ruby test source, the test framework, the test documentation, the operating system source, and the OS documentation would be several feet tall.
A working demonstration of an electronic visa application is complete. It demonstrates the use of Caernarvon access controls to permit authorized countries to read biometric data from and write entry/exit time stamps on each other’s passports. It also demonstrates how the same access controls permit unauthorized countries to read public data but prevent them from writing anything. Lastly, it demonstrates how initialization data can be permanently write-protected from all countries.
"The physical level will not indicate presence and internal workings, because the hardware runs off a "constant current generator", and the hardware contains measures against DPA."
Tells me you are actually not that familier with the nature of the problem.
The mear fact that there is a tuned circuit with a load on it indicates it's "presence" in exactly the same way that a lot of store "anti-theft" tags work it is an unavoidable charecteristic. It can be charecterised irespective of what measures the chip takes to prevent faulty operation (as these protective measures are charecteristics in their own right).
Secondly prior to the electronics in the chip getting to a stable state it is not possible to mask the activities going on on the chip. However each chip "mask" will have a recognisble power signiture in this power up phase so it will be possible to recognise chip types and stepping of the "mask" used to make them.
This will in all likleyhood reveal which country issued the passport and aproximatly when, therfore also aproximatly how long it will be valid for.
For some low technology forms of attack this is more than sufficient, when you can assess the person holding it with your naked eye...
With regards to high technology attacks, systems are made to a price, in high volume activities usually the bottom price gets the job (such is the reality of transport cards, phone cards, credit cards etc).
When a low price has to be achived the functionality tends to be limited and embeded sensors etc tend to have high thresholds to maintain production yields.
Also the design of the current passports is such that the chip is relativly easy to get at and compleatly under the end users control once issued.
Further to this even when the chip is powered up there are ways of injecting signals into the chip and determaning it's functioning and active states.
One such method is to subject it to a low power plane wave microwave signal. This will effectivly be re-modulated with the activity on the chip. Re-mixing this resulting signal with quadrature waveforms of the original source will result in a base band signal that can be analysed in the same way as DPA could. Unlike DPA which is essentialy baseband to start with you have an extra degree of freedom in that you can very the frequency and power level of the microwave source.
Importantly the level of signal required to do this is less than the threshold to cause unexpected activity that can be easily detected by on chip sensors.
Similar attacks can be performed using magnetic fields from nano probes laser diodes with frequency up converter crystals and even electron microscopes.
Experiance tells us that for a determined adversery there will be successfull avenues of investigation that more often than not result in exploitable avenues of further investigation.
And as Bruce has noted attacks only improve with time, and the ten year life expectancy of a passport is in electonic terms nearly seven generations.
So, technically it is feasible from the hardware side, and already available.
From the software side it is to be included in a standard by ISO and is in the process.
The infrastructure will just need a little software update to run the protocol.
I think the western hemisphere std with the flaws is something else? Sounds like crap.
I agree that you have the resonant circuit detectable. In a correct implementation an analog circuitry draws a constant current from the resonant circuit, essentially an analog current stabilizer. As long as this is not satisfied with the current, it keeps the digital part silent. Better smartcards contain such sensors.
The mask, stepping etc will not tell you anything. What you have to do is to get at the "private key", which is on the card and different for every card. This is only possible by physical analysis of the chip.
This "private key" belongs to the "public key", which is bound to the remaining information by a signature of the issuer.
So the public key, together with name,picture,fingerprint etc. cannot be separated to create a new identity, without having the private key of the issuer, e.g. the federal printing office.
In the result: You need high-tech physical analysis to just clone a card, without changing any information, name, picture, fingerprint etc., because that is the only way to get the individual private key of the card.
This private key of the card is not good for construction of any new identity, because it is linked to the public key and thereby bound together with the information of the card.
So high tech analysis can not help you to corrupt the system.
You need the private key of the issuer to do that.
One last thought for you to consider.
Even if the system could be made unbreakable will it be accepted?
For instance the SET protocol failed because it offered the customer a worse case than they already had.
Likewise other systems like electronic purses where rejected for similar reasons.
Sometimes people have less than obvious reasons for rejecting technological improvments.
Apart from cost I can think of several reasons why a lot of countries would not want unforgable passports and visa systems implemented, esspecialy with audit trails.
What they want is a system sufficiently complex to be unforgable without sufficient technical resources. But still forgable to their own organisations that do have the resources.
An example organisation would be Mossad of whom it has been publicaly stated that they routienly forge passports of other nations.
Nation States will always want to quietly play their little games against each other. What they don't want is to be politicaly embaressed by extreamist / terrorist events happening very publicaly in the streets of their captial cities. Or the press hammering on about economic migrants stealing peoples jobs.
As we now know passports are not an issue for extreamists / terrorists the current crop simply find their way in by boat etc on insufficiently protected boarders, or come in on their own passports.
Likewise proffessional criminals are not that worried by passports.
And governments are actually not that worried about economic migrants as long as they can tax them or deny them access to resources the government funds.
In fact in many countries economic migrants are activly wanted for many reasons, not just to fill tempoary workforce shortages but also to stop wages spiraling out of control.
Well, then the Mossad has to hurry up their operations a bit with the new system, before the chopped off finger reaches its expiration date... And actually for that case there is also the concept of dual citizenship, so why would they worry about passports too much. Regarding the economic migrants I don't see the connection, but I am biased because I am one myself. What the described solution prevents, is that in some kitchen, shop floor, printing office and photo lab etc. somebody with limited means takes blank passports, invents an identity, adds fingerprint and photo and has a valid passport. That is already better than with some traditional methods like watermarks, holograms. Also it removes the factor of deniability. If a country-X electronic passport shows up and the signature checks out, then you know that it has been issued by their federal printing office, or their federal printing office has given out keys or been bribed. So they know whom to shoot or hang.
And finally, there might be applications, where you have to authenticate yourself as the operator of the federal computer, nuke silo, bomber hangar, plutonium facility etc., and you also there don't want somebody to glue his photos into a blank access card and sign them as "Donald Duck"... So there is enough technical motivation to pursue the technology. And the described system technically doesn't sound bad, does it?
"The mask, stepping etc will not tell you anything. What you have to do is to get at the "private key","
For a technological attack that is true but that is not the point.
Technological attacks such as cloning are high hanging fruit, mugging and identity theft are usually not technological in nature and are usually low hanging fruit.
For ID shopping to get into a country all I need is the passport and fingerprints of a person from that country that sufficiently resembles me.
With my own eyes or others working for me finding a backpacker who looks sufficiently physicaly like me is not going to be that difficult.
Finding out if they come from the right country can be found from the chip mask setting. And if there is sufficient need then a pocket device could be made to do so.
I simply then wait for my target to be alone in an apropriate place and mug, kill or kidnap them. I can then if required duplicate their fingerprints in a quite low tech manner sufficient to pass most "currently in use" fingerprint scanners and I then make myself look like target sufficiently to pass visual examination.
I then have a short time window to fly back to their country where I simply disapear into the background as simply as any economic migrant currently does.
And as most countries don't currently scan fingerprints or electronic passports currently, and as I have found by accidently traveling on a friends passport check your apperance with that in the passport you have a reasonable chance of getting away with it.
99 times out of a hundred hi tech fails to low tech attacks based on exploytable human weakness, and I don't expect that to change any time soon.
Disappearing into the background could take on a whole new meaning with your above mentioned new friends!
I certainly do not understand, why the discussion about migrant workers and economic migrants is coming up in the context of the electronic passport.
I am a migrant myself, and I take away your job, you bet! But my documents are ok.
Besides, in the described system the passport programming machine in the federal print office also generates its key pair internally and gives only out the public key. So fortunately no human operator can get at the private key without doing physical analysis of this machine. (destructive physical analysis, to be precise)
I have one of the WHTI cards (electronic passport) that can be used at border crossings (land, sea & airports). It comes with a shielded envelope so that it can not be read while stored within. The only time I take it out is when I am in a secure Customs & Border Protection area in an airport for a few seconds while using a CBP kiosk to avoid the long lines for those who don't have one of these cards. I am hoping that my risk is minimal given I am in a secure area and that someone trying to sniff this info during the few seconds I have the card outside the envelope would have difficulty hiding their capture gear in a location filled with security agents. Am I being overly naive about the risk during this limited exposure of my card?
"Am I being overly naive about the risk during this limited exposure of my card?"
There's four basic things involved here,
1, Is an attack possible.
2, Is anyone using an attack.
3, Is there an effective defence.
4, Are you being targeted.
There is also the unknown elerment of changes with time due to other factors.
The first question is fairly easily answered in that yes there are a number of attacks possible and I would think it likley that more will be discovered with time (this is the norm for new classes of security systems).
The second is currently fairly easy to answer as well. The system is "to new" and there are "to few" cards out there currently. And as with most types of activity there is the question of "to what advantage" and currently there are easier ways of carrying out most ID related crimes. So at the moment apart from researchers the answer is most probably no.
However with time more attacks will be discovered some of them very low tech some high tech. Then the question of "to what advantage" comes into play.
If ID theft for ordinary criminal gain becomes more prevelant then the principle of "low hanging fruit" will be applicable. That is as security increases in other areas attackers will move to where the "sweet spot is". If for instance mission creep happens and these cards are required for opening a bank account or for other financial transactions then they are going to become targets.
Also there is the question of "how high the bar is set". We have seen with online fraud what is effectivly an arms race between attackers and defenders. If the Banks etc had introduced effective levels of security from day one it is considerably less likley that we would have the curent sophisticated attacks or levels.
Which brings us around to defensive measures. The information as presented is not unexpectedly that shielding is only partialy effective. And the more convienient for the user the less effective sheilding is.
Again there is a time issue here, like anything mechanical shielding is subject to wear with use and ultimately failure. To see this take a piece of solid metal wire and bend it a few times, it cracks and breaks.
There is of course one defensive measure that works 100% against the sort of attacks being discussed here. If you either do not have one
of these contactless devices there is nothing to attack.
However at slightly less than 100% is if you keep it locked up in a suitable "safe" then it is not going to be possible to attack it. So the simplest and most effective measure if you have one would be to only carry it when needed to minimise exposure time.
Although the down side is aproaching the place where you have to use it as there is a reasonable expectation you will be carrying it. And if the cards can as appears to be the case reliably read from 100ft away then yes there is the problem of using it for it's intended purpose, which is where an attacker is most likley to be if the Boarder Patrol ordinary security is not up to the job...
Which brings us around to "are you a target for attack". To which there are the usuall answers to any type of crime "means, motive and opportunity".
The means has been discussed and as it is possible I suspect that at some point it will be used. However how wide spread it will become is another currently unanserable question.
As to motive if you have what a criminal wants then you are more likely to be attacked. Jewel thieves generally target those they know have what they want not randomly "jo average". If there is sufficient motive then you will be attacked.
However crimes like mugging and credit card skimming are crimes of oportunity and to most people are considered "random attacks".
However they usually are not random which is why some neighbourhoods are "no go areas" for those that can avoid them. And why some of those that cann't take protective measures.
The important thing is what the armed forces refere to as "situational awareness". Generally most people get hit when they feel at ease or do something that makes them a target. If you are aware of this then you can take preventative measures to protect yourself.
However remember the measures should be proportianate to your risk level not somebody elses. Which is the rub as currently we have no information on what the risk level is as it's all "to new" and might or might not change with time...
So for now as you have one leave it at home in the safe untill you need it and then carry it in it's shielded wallet and optionaly wrap it in tin foil as well to improve the shielding.
If the chip transmits at 13.5MHz, couldn't we just build a portable transmitter operating at the same frequency to jam the signals within, what, 100 feet?
"Researchers at the University of Birmingham, UK, have shown how ``e-passports allow real-time tracking"
I generaly find the best place to drop such news is on Bruce's friday squid page ;)
Here is a link to the paper,
It appears to be a more restricted but more reliable version of the card "fingerprinting" I have talked about in the past.
The hard part as they say is getting the query from listening in to an authorised reader. The question is are they going on to find out about "test/engineering codes" which would obviate the difficult step.
Also the attack works at a greater distance than 20 inches with the right type of loop.
Do you want to re-post to the friday squid page?
20inches seems very short! I presume better equipment would enable this distance to be increased?
I'll repost to Friday's squid post, once the Squid raises it's head from the water...
"20inches seems very short! I presume better equipment would enable this distance to be increased?"
Yup, but that's the figure they give in the paper etc.
At 13.5 MHz you could but a loop around the door frame with little difficulty and if you know what you are doing make a linear GTem cell.
But hey I don't think range was realy on their test agender (as others have done some work in that area).
There is also the issue of the RFID coil alignment etc so they may have ment 20in minimum.
Schneier.com is a personal website. Opinions expressed are not necessarily those of BT.