Arjen's blog

London is your Oyster

The Oystercard is a electronic debit card that has all but replaced tickets on the London tube and bus transport systems. It allows users to put money on the card and discounts this credit as the card is used to enter and exit the underground and buses. The system is fast and unobtrusive and almost everyone who uses London public transport has one. Everyday millions of pounds are being put on these cards and taken of again as Londoners move about their city. The security mechanisms that are supposed to be keeping these millions (your money ultimately!) safe have now been shown to be, frankly, utter crap.

So the Oystercard has been definitely compromised. Its fundamentally broken and needs to be replaced by new technology concepts. Independent experts warned about this back in 2004 so there's no excuses for being all surprised now. Then they warned again in 2005 and last December at the Chaos Computer Club  international IT conference in Berlin. Now the attack on the lackluster security system of the London Oystercard has been practically demonstrated last April by a group of Dutch researchers who were investigating the same technology that was about to be implemented nationally in the Netherlands. "The Oystercard system uses the same chip and has the same basic vulnerabilities" according to Professor Bart Jacobs of the Computer Science faculty of Nijmegen University.

After the publication of some of the inner workings of the data-encryption mechanism of the chip used in the Oystercard last December in Berlin, a fully operational breach was predicted by many experts. With the basic knowledge of the inner workings of the chip available online for anyone to see, implementing a working attack against the system was just a matter of time. The Dutch research group has been able to clone the funds on an Oystercard to another Oystercard. This provides at-home top-up mechanism allowing essentially free travel in the greater London area after an initial investment of 10 pounds plus a few blank Oystercards at 3 pounds each. Since the required devices and software are otherwise pretty much free today or in the near future (at most a few months from now), the London Transport Authority needs to get moving on this or accept that they will be providing free travel for those capable of using a laptop and a high-end mobile phone (and all their friends).

The systems failure bears all the classical hallmarks of public-sector IT screw-ups. Basing your security mechanism on trying to keep the inner workings of such a system a secret while at the same time distributing 12 million copies of said system into the hands of the public is, frankly, insane. Does anyone think a handful of engineers locked in a room at Philips can come up with a system clever enough so that the combined expertise of 1 billion Internet users cannot defeat it? One has to wonder what they were smoking that day. Then there are bonus points for ignoring repeated warnings from independent experts for several years.

Among security professionals it is considered scripture that the only systems that can be trusted are those that have been tempered in the fire of public scrutiny. No one is as clever as everyone, and with a few million interested specialists online there is nowhere to hide for a system containing design flaws. Flaws are always found sooner or later and most often sooner. One would think that after six decades of spectacular failures the method of keeping a system secure by trying to hide its inner workings (know in the security trade as 'security by obscurity') would be utterly invalidated. The Germans used this method for their supposedly secure communications using the Enigma machine in World War 2. It cost them the battle for the Atlantic and ultimately the war (ok, ok, attacking Russia in August without winter coats for the troops was not a smart move either). More recently a $400 million DVD encryption method was broken by a 16 year old Norwegian hobbyist.

The Dutch system will probably not be implemented in its current form, but the London system is already operational with an estimated 12 million people using the card. For the sake of the financial stability of the tube system one can only hope that clever engineers have already been working on a solution that can be implemented quickly. But I'm not holding my breath.

(earlier post on the Dutch card here - posts in Dutch here)

The Times picked up the story, and has a writeup of the wider security implications (accescards to buildings and such).

Bruce Schneier on security trade-offs

Bruce Schneier just posted a really good explanation about "the difference between feeling and reality in security". It is one of those articles I wish I'd written. Not because there is a great new nugget of insight in it but because it explains some very basic problems in thinking about security so very well.

The gist of the article is that as people living in modern environments we can have a hard time accurately estimating realistic trade offs between risks and reward. When the world was closely resembling the world we had developed in as a species we were better at it. Our brains were supported by millions of years of evolution in correctly estimating the risk versus rewards of certain actions. There's food here and a lion, should I stay or run? The specimens who made bad trade-off calls died of hunger or lions. The ones making good calls had many babies.

The risk vs reward trade-off estimations have become a lot more complex in a hightech, globalised world. How much should we spend of keeping us safe from terrorism versus say, tripping on our wet bathroom floors? The fact that other parties have political or financial interest in making it harder for us to make good choices in these matter complicates things even further. As with many things critical thinking and good information are key to making better decisions. And don't worry, today those do not necessarily lead to many babies.

More of Bruce Scheiers thinking van be found on his blog and in his 2003 book 'Beyond fear' that can be read about here, read online here entirely and bought in dead-tree form here. I highly recommend it to anyone who needs to think about security issues, not just information security.

Get a famous fingerprint

The German Chaos Computer Club, the oldest and largest hacker group of Europe, made available to the public the fingerprint of the German Minister Schäuble for the Interior. They wanted to show how easy it is to obtain someone's identity when identity is based on fingerprints.

The German government is preparing to build a national database containing the fingerprints of all its citizens for the purposes of fraud-prevention and national security. Minister Schäuble is very angry about the release of his fingerprints and has stated he will take legal measures against the CCC. Dutch hacker Rop Gongrijp pointed out that the Minister's anger was curious since it was the minister after all who wanted to collect the fingerprints of over 82 million Germans and the CCC only collected one.

The CCC has been demonstrating for several years how easy it is to 'steal' someone's fingerprint and use is to fool all kinds of security measures such as payment systems, physical access controls and computer security systems. As with the doomed RFID cards these demonstrations need to be very 'in your face' before media and governments take notice. Worldwide there are over 200 million devices in use of the 20 different types that were fooled by the CCC experts. As with the 100 million RFID cards they are all essentially worthless as serious methods for securing transactions or granting access.

It is curious how we as citizens are constantly required to trust governments to handle our most private data when these governments often are not that trustworthy themselves and also not very technically competent in guarding our information. Passports are easy to fake, RFID cards are easy to copy, fingerprint readers can be fooled. Before we base our entire lives on these technologies we'd better make sure they actually provide a minimum level of security. For now I'm sticking to encrypted mail and strong passwords.

German TV broadcast an item about the possibility of stealing a fingerprint and using it to go shopping at someone else's expense at a large German supermarket chain.

Since the TV piece did not include the entire method for making your own fingerprints I include it here. As with the RFID cards, these vulnerabilities have been known for several years, it's just that some companies and governments are a bit slow in picking up on them. If you want to go shopping as Minister Schäuble, just click on the picture at the top and follow the procedure from the movie.

A friend and IT security expert pointed out that since anyone can now pretend to be Minister Schäuble, that pretty much makes his fingerprint useless as evidence in  court. Maybe we should all publish our fingerprints (and retina scans and DNA profiles) to gain plausible deniability on future accusations of anything ...

Public Transport card fully hacked

What experts foresaw last December and the Dutch research institute TNO denies was possible in their recent report has been done. The deepest level of data-encryption on the NXP Mifare RFID chip has been hacked. Cash from cards can now be copied to other cards through cloning and that makes this system utterly unsuitable for serious applications involving real people and real money.

But this is essentially old news. The more interesting news as far as I'm concerned is the fact that TNO was immediately re-hired by the company implementing the card system to do more research on the validity of the hack. You have to wonder what the thinking is here. This company dropped the ball on at least three separate occasions in this area so why do they get another chance to write a big rapport to claim 'there is no problem' ? And this is not the first time, on the sensitive subject of voting computers (now banned in The Netherlands) they also kept telling us 'all is well'.

If you merely want a paper to reassure yourself just ask the secretary to print out a pretty picture from the Interweb with a caption that says 'everything will be ok'. That's a lot cheaper then hiring a company like TNO, and apparently just as valid.

Does TNO just write down whatever the customer asks of them or do they really not know any better? Either alternative is troublesome. As an important expert-adviser to the governments we should hold TNO to a higher standard. When faced with an impossible request from a client they should respectfully decline the job explain that the client's request is either technically impossible or not in line with laws concerning citizen privacy and such.

Our government (and parliament!) allowing such organizations to indirectly guide technology policy is a real problem that will continue to cost us dearly (in real money, privacy violations, theft and missed technology opportunities).

Next up for big IT-projects is a road-toll system that should allow for more flexible costs of (pay-as-you-go) owning and using a car. Hopefully it won't be as insecure as this project, that could be expensive for government or the citizen (or both).

Tomorrow there will be a meeting in parliament on this matter. The independent experts (the ones that got it right from day 1) have decided to boycott this meeting since they are not allowed access to the 'secret' paragraph of the most recent TNO rapport. They wisely refuse to legitimize more 'security-by-obscurity' bullshit.