Hacker News Reader

Comment by 283894 | Hacker News Reader

I just signed up for 2 Tesco accounts the other day to dump 3k in each for the 3% interest.

I'm certainly not going to be doing anything with the accounts until Tesco give some more clarification on what actually happened (although the way these things work, I doubt there will ever be a full technical response.)

Also if it is some sort of internal breach, would any other data have been taken?

Back in 2012, Tesco were storing passwords in plain text.

http://www.bbc.co.uk/news/technology-19316825

View on HN

Tesco ask you to log on with "character 2 and 4 from your password" which sort of implies they must store the password in clear text (unless some kind of zeroknowledge/homomorphic encryption magic i've not heard of.)

That is not the password though. They call that the 'Security Number'.

After entering the two digits of the 'Security Number' you then receive a 'One Time Access Code' through a text or phone call, although I have never logged in to my account before, and seem to be unable to get past this step now.

I think you then enter your proper password in, which I would hope is not stored in plain text, although the article I linked seemed to imply this was the case back in 2012.

Depending on the length of the password, it's possible to encode/hash (+salt) all possible outputs of challenge combinations at the point of storing your password.

It's a bit like having a number of related passwords, which the bank can ask you for any of them, and then verify is correct.

I have an account at Metro Bank in UK. One day I was on the phone with them, and to authenticate myself they asked me for characters 2/4/7 from the password.

At the same time, they advise you to never give your password away, and that they will never ask you for your (full) password.

Talk about a mixed message...

Most UK Banks use HSM modules to store 'pins' (Similar process to Apple with iCloud) which remain separate to passwords which are hash + salted

This is pretty standard for UK banks.

They'd do a hash of each character of the password (in Lloyds' case, your "memorable word" combo), to compare your entries to.

Wait, that's still awful! It allows you to crack each character individually. For instance, a 10-letter password requires 26^10 ~= 1.4e14 attempts to test every option if you only have a hash of the full password, but only 10*26 = 260 attempts to test every option for every individual character.

I have to use two passwords to login to Lloyds bank. One conventional password (which is presumably stored salted and hashed) and one where I have to enter characters from three positions they choose. The latter is intended to mitigate the risk of using your account from a vulnerable computer. The former takes care of vulnerabilities on their end (as far as any password can).

Could they implement something like:

Password: money

Secret word: ABCD

If they're going to ask for two characters from the secret word, they could then hash

  saltmoneyAB
  saltmoneyAC
  saltmoneyAD
  saltmoneyBC
  saltmoneyBD
  saltmoneyCD

and check against the relevant one.

I may be incorrect on the "hash each letter individually" part. But this is combined with a password.

Reply on news.ycombinator.com