NO2ID

'The gold standard security algorithm that underpins online transactions has been broken by researchers'

It will take five years to develop another but don't worry for even the most serious hackers haven't got the computing power nessecery to do it, YET! No doubt Mr Blair and Mr Clarke will take this into consideration when introducing the NIR.

Won't they?

Im not sure what you mean by ur post Justin!
is this a form of security structure for on line transactions that has been broken & made redundant?

As I understand it, yes.

All online transactions require encryption and it has been assumed up until now that the alogorithms being used to encode infomation were unbreakable due the the fantastically large number of possible 'codes' they generated. Now that these algorithms have been cracked in the lab using the latest supercomputers it is a reasonable assumption that they may well be broken by less savoury types within a few years. No transaction via email can then be guaranteed secure, including of course the passing of information that resides in the proposed central NIR whilst checking ID cards.

In other words the information stored on us will be available to unauthorised hackers and this will include any credit card details as well.

If quantum computing ever makes it out of the lab then all our attempts at encryption will be rendered obsolete anyway due to its incredible computing power.

Justin.

I knew many "good" coders over the years but what scared me was this twelve year old who could crack 128Bit and more encryptions when they were cutting edge...

A team of hackers *could* break this code quicker if they collaborated such as people do with seti @ home which divides the power needed by the users helping, another example of this would be the protein "folding" crew that works to fold and unfold protein chains to help in diagnosis and cancer research.

If hypertheticaly they needed to break the codes, what would they need? a warehouse with a second-hand mainframe maybe? Or a series of twin proc'd Xeon's running segments of the cracking program? People do tend to think in single use PC terms but there are some phenomally powerful machines out there up for grabs.

It still comes down to the "what can be made can be broken.." principles and I have kept up with this 12 year old who is in his early twenties, if he ever did turn his mind to criminal activities then he would be pretty unstoppable.

Do you have any more information about this - this article doesn't appear to be on their website and I haven't seen anything about this in either the Information Security press or the security mailing lists or the underground mailing lists I read.

The only algorithm that has been defeated, that I know of, recently is SHA-1. Which is a hashing algorithm - not encryption.

Ancientone: If you really wanted to crack encrypted traffic, you'd need:

1. A large selection of the encrypted traffic (could be got by sniffing in the right place).
2. A large grid computer.

One solitary mainframe would be pretty useless, just like cracking on your solitary PC would be - mainframes are fast and powerful because they have the hardware designed around the fast moving of internal data. In terms of raw calculation (which is what you'd need for attempting to crack encryption algorithms) you need CPU power.

When the EFF won the challenge to crack DES, which they could do in 23 hours, they built a special grid computer with custom chips designed to just process DES.

Note, the difference between the number of different keys in 56-bit DES and 128-bit RC4 (the current favourite algorithm) is to the power of 56.

56-bit DES had a maximum of 72057594037927936 different keys, whereas 128-bit RC4 has a maximum of 3.40282x10^38 (i.e. a 38 digit number).

This is an astronomical number, so to put it in perspective; if we say that you could calculate 1,000,000 keys a second and run this on a million nodes of a grid, this would still take 3.40282x10^26 seconds to crack one piece of encrypted data. This is equivalent to 107902714358193810248 years.

Without a flaw/backdoor in the algorithm or an increase in computer power well above Moore's Law this is not going to change.

I'm intrigued about your 12 year old hacker - again do you have any details - which algorithms and what techniques did he use?

From the Telegraph 8/2/05



Thanks r.d for a succinct piece. It is however true the the technology is continuing to move at speed (as above).

One can only hope that the encryption technology keeps with that of the computers used to crack it.

The weakness as I see it is not in the data traffic once it has been encrypted, but in the point-of-service themselves. Beacuse there are many thousands of these they tend to be manufactured "down to a price" using the cheapest (least powerful) components. If they proved vulnerable ....

Should read "point-of-service terminals", sorry.

Cryptanalysis of SHA-1
http://www.schneier.com/blog/archives/2 ... sis_o.html

The published theoretical weakness in the standard secure hash function alogorithm which is used for Digital Signatures e.g. to make data like Biometric Identifiers stored in a Smart Card secure against tampering or forgery is big news.

The exisiting alogrithm is not, for practical day to day purposes broken at the moment, but the "would take all computers on the planet several lifetimes to crack" comfort factor that cryptographers and security experts demand is no longer true.

The feeling seems to be that SHA-1 should be replaced within 5 to 10 years, which is ok for relatively trivial problems like sending billions of pounds of money around the world electronically, but that is not good enough for a "gold standard" National Identity scheme, where the system needs to be secure and unforgeable for your whole life from the age of 16 and for several years after your death (to prevent criminals stealing dead person's identities a la Day of the Jackal)

Assume that the first issue of UK ID Cards uses SHA-1 and by the time they expire in 5 years time (N.B. no smart card manufacturer promises a longer guarantee than this, not necessarily due to failures of the electronics, but wear and tear on the plastic cards themselves, a factor which effectively doubles the cost of what is currently a 10 year passport)
computing power is deemed to be sufficient to introduce say SHA-256 (if they can be sure that a longer more complicated secure hash algorthm is actually orders of magnitude more secure).

What happens to the audit trails and the actual records on the National Identity Register that were signed with now broken or unsafe SHA-1 algorithm ? These are long term credentials, and there is no experience or technology of handling these securely on the scale that the NIR would involve.

How could the system cope with, for instance , an ex-patriot emigree who is returining to the UK after living overseas for the last 5 years ? If his or her old ID Card was signed with the "broken" SHa-1. Later when applying for a new one, using the old ID card as one of the supporting documents (along with the other trivially forged "utitlity bills" etc.), then how can the system be sure that the ID card is not a forgery or that the person presenting the old ID Card has not subtituted their own Biometric Identifiers onto the card and digitally signed them using the "broken" SHA-1 algortithm i.e. "Identity Theft" - exactly what the whole system is alleged to prevent in the first place ?

Storing complete sets of raw image /scan data of fingerprints, irises, facial photos etc. is impossible on current Smart Cards (it is possible on some optical storage cards, but these are less physically tamperproof) and an insane idea if they are transmitted from the Registration offices to a Central Database. The risk of a corrupt or blackmailed or treacherous "insider" having access to the raw biometric scans of people means that they can "mint" their own ID cards much more easily, than if, as I suspect, the designers of the NIR simply store hash function based checksums of samples of the raw data i.e. "iris codes", "fingerprint miniutia" etc. which are reasonably ok for comparing with those stored in an ID card or from a Biometric Reader when the person uses the system.

Lifetime (plus a few years after death) long term secure electronic credentials, with all the problems and risks of changing cryptographic algorithms several times, have never been done on the scale that the UK Government seems to be hoping for can be achieved through "technological magic".

I think the Government may be even more technologically (and otherwise) naive than you suggest. "Security" is presupposed for the whole system as if it were a trivial detail that could be sorted out later, and mostly covered by deterrance.

The mind-set of the Home Office (and most other departments) is that if you make something an offense and provide a stiff penalty, you will get compliance. They cannot comprehend anyone breaking the system out of curiosity, pique or simple unwillingness to comply. They cannot comprehend probes, or serial, adaptive attacks. They assume that an attacker will be a deliberately malevolent one-off; and that everyone else has nothing better to do with their lives than follow rules.

Arbitrary enforcement powers are likewise (for all habitual waffle about "striking a balance") sought for convenience--to avoid the possibility of challenge, not because of any real calculation about effectiveness outweighing justice.

I keep in contact with him from time to time... When he nexts contacts I will ask him but this is like witchcraft to me so not fully in the position of being able to comprehend it all in entirety.

I notice on some of the P2P sites a sudden release of next gen credit card generators, (I use P2P so I can get access to Alex Jones/Naom Chomsky and other documnetaries and files so before I get accused of piracy etc. I don't - There are legitimate uses for P2P.)

I am not going to download one of these but there is a buzz that these CC gens are the real business and use the correct algo's and hashes that are in place for online transactions.

If this is the case then it may be that the hashing system is already broken to a degree that a software program can reproduce and create the correct information responses that work with the online transaction services, I'll try and find some more out on these as well but bear in mind that its a really dodgy area and I will only go so far to elucidate any info.

If however that a heavy duty system like SHA1 has been busted then it shows again that the government could not and cannot rely on this as a security system.

As for the other things, one of the things I proposed was the creation of a closed network such as seti @ home, there are loads of these projects out there, it may be a false flag op could be used to dupe PC owners to assimilate some of the data processing under a false pretence, certainly many underground and piracy download sites seemed interested in this "folding" thing.

If I remember aright though about data streaming and capture, this is usually done by a planted software or hardware bypass that echoes the stream or diverts it through an unspecified route, the apparent ease that the US security services have in implenting this without the ISP's awareness would suggest that is not as hard if you have the tech and the resources and there are many criminal gangs out there on the net now who fleece the system for money and information on a well tried and regular basis.

I will report back when my fried contacts and I will leave this subject to the experts... I will say no more

You won't get me arguing there - the chasing down by the RIAA, MPAA and local pawns is one of the scarier things happening on the Internet at the moment...



Oh, the credit card generators are real. The problem here is that we are comparing apples with pineapples. A credit card number is quite easy to generate, especially as the check digit is only based on a very simple algorithm.

Whereas, a general hashing algorithm is used to make a cryptographic hash of data. Note, this will not be able to be reverse engineered back to the cleartext, unless the cleartext is smaller than the hash (and it still requires a large amount of processing power). The hash value of, say a fingerprint, would probably be sent to the server.

The 'broken'-ness of SHA1 (just like SHA0 and MD5 beforehand) means that the researchers have a way of producing collisions in the hash. Whether it can be abused depends on how the system is implemented - can I inject a hash value in at certain points?

Cryptographically, the fact that the algorithm has been broken means that it shouldn't be used.



That is the biggest risk - spoofing of another individuals ID information.



Or the bot-nets (virus infected PCs - currently being used to send out spam emails) or the 'chinese radio lottery' concept.

The problem here is that though grids do provide a large amount of computing power they still don't have enough to perform custom cracking of encryption - generally specialist accelerators are used in these situations.



Oh; it's a lot easier than that - a tap or span on the routers will allow somebody to passively copy everything on the network, and these devices may be made totally invisible to network (only the switch/router providing the span/tap will be able to see the device).

What we have to be careful of is getting too far down the 'tin hat brigade' area and making sure that our arguments are based in fact, and not sounding paranoid.