SSH an ill-managed mess says SSH author Tatu Ylonen
IETF draft a first step to new version
Tatu Ylonen, author of the SSH protocol, isn't afraid of criticising his own work: he's calling for a new version of the Secure Shell to make it more manageable and get rid of the problem of undocumented rogue keys.
In this IETF Draft, Ylonen proposes a regime for key management, including key discovery, to overcome the problem. The draft, co-authored by NIST's Murugiah Souppaya and Secure IT's Greg Kent, proposes guidelines for “discovering, remediating, and continuously managing SSH user keys and other authentication credentials”.
The draft notes that there are often a great many more SSH keys in existence in an organisation than there are users – “hundreds of thousands, even over a million SSH keys authorising access have been found … [in] many large organizations. This is many times more than they have interactive users.”
Of course, having credentials wandering around loose in a company “present a real risk to information security”, the draft notes – and places a premium on finding out just what's out there.
This draft, open for comment until October, focuses on processes and key management, but according to Ylonen, a new SSH will be needed (the current version, SSH-2, dates from 2006).
Some of the recommendations in the draft appear to reveal a surprising state of affairs, at least among companies whose SSH implementations have been reviewed by Ylonen's company, SSH Security. The draft recommends:
- Moving keys to protected locations;
- Removing unused keys;
- Associating authorised keys with a business process or application;
- Removing keys for which no valid purpose can be found;
- Rotating keys;
- Restricting what can be done with each authorised key;
- Establishing a process for approving new keys.
The list looks somewhat like what companies should already have been doing in managing their SSH access – but it's probable that someone configuring new kit finds it simpler to create such things ad-hoc. ®
Sponsored: Hyper-scale data management