/mandos/trunk

-*- org -*-

* Mandos

  - Have your cake and eat it too!

  You know how it is.  You've heard of it happening.  The Man comes

  and takes away your servers, your friends' servers, the servers of

  everybody in the same hosting facility. The servers of their

  neighbors, and their neighbors' friends.  The servers of people who

  owe them money.  And like *that*, they're gone.  And you doubt

  you'll ever see them again.

  That is why your servers have encrypted root file systems.  However,

  there's a downside.  There's no going around it: rebooting is a

  pain.  Dragging out that rarely-used keyboard and screen and

  unraveling cables behind your servers to plug them in to type in

  that password is messy, especially if you have many servers.  There

  are some people who do clever things like using serial line consoles

  and daisy-chain it to the next server, and keep all the servers

  connected in a ring with serial cables, which will work, if your

  servers are physically close enough.  There are also other

  out-of-band management solutions, but with *all* these, you still

  have to be on hand and manually type in the password at boot time.

  Otherwise the server just sits there, waiting for a password.

  Wouldn't it be great if you could have the security of encrypted

  root file systems and still have servers that could boot up

  automatically if there was a short power outage while you were

  asleep?  That you could reboot at will, without having someone run

  over to the server to type in the password?

  Well, with Mandos, you (almost) can!  The gain in convenience will

  only be offset by a small loss in security.  The setup is as

  follows:

  The server will still have its encrypted root file system.  The

  password to this file system will be stored on another computer

  (henceforth known as the Mandos server) on the same local network.

  The password will *not* be stored in plaintext, but encrypted with

  OpenPGP.  To decrypt this password, a key is needed.  This key (the

  Mandos client key) will not be stored there, but back on the

  original server (henceforth known as the Mandos client) in the

  initial RAM disk image.  Oh, and all network Mandos client/server

  communications will be encrypted, using TLS (SSL).

  So, at boot time, the Mandos client will ask for its encrypted data

  over the network, decrypt it to get the password, use it to decrypt

  the root file, and continue booting.

  Now, of course the initial RAM disk image is not on the encrypted

  root file system, so anyone who had physical access could take the

  server offline and read the disk with their own tools to get the

  authentication keys used by a client.  *But*, by then the Mandos

  server should notice that the original server has been offline for

  too long, and will no longer give out the encrypted key.  The timing

  here is the only real weak point, and the method, frequency and

  timeout of the server’s checking can be adjusted to any desired

  level of paranoia

  (The encrypted keys on the Mandos server is on its normal file

  system, so those are safe, provided the root file system of *that*

  server is encrypted.)

* FAQ - couldn’t the security be defeated by...

** Grabbing the Mandos client key from the initrd *really quickly*?

   This, as mentioned above, is the only real weak point.  But if you

   set the timing values tight enough, this will be really difficult

   to do.  An attacker would have to physically disassemble the client

   computer, extract the key from the initial RAM disk image, and then

   connect to a *still online* Mandos server to get the encrypted key,

   all *before* the Mandos server timeout kicks in and the Mandos

   server refuses to give out the key to anyone.

   Now, as the typical SOP seems to be to barge in and turn off and

   grab *all* computers, to maybe look at them months later, this is

   not likely.  If someone does that, the whole system *will* lock

   itself up completely, since Mandos servers are no longer running.

   For sophisticated attackers who *could* do such a thing, *and* had

   physical access to the server for enough time, it would be simpler

   to get a key for an encrypted file system by using hardware memory

   scanners and reading it right off the memory bus.

** Replay attacks?

   Nope, the network stuff is all done over TLS, which provides

   protection against that.

** Man-in-the-middle?

   No.  The server only gives out the passwords to clients which have

   *in the TLS handshake* proven that they do indeed hold the OpenPGP

   private key corresponding to that client.

** Physically grabbing the Mandos server computer?

   You could protect *that* computer the old-fashioned way, with a

   must-type-in-the-password-at-boot method.  Or you could have two

   computers be the Mandos server for each other.

   Multiple Mandos servers can coexist on a network without any

   trouble.  They do not clash, and clients will try all available

   servers.  This means that if just one reboots then the other can

   bring it back up, but if both reboots at the same time they will

   stay down until someone types in the password on one of them.

** Faking ping replies?

   The default for the server is to use "fping", the replies to which

   could be faked to eliminate the timeout.  But this could easily be

   changed to any shell command, with any security measures you like.

   It could, for instance, be changed to an SSH command with strict

   keychecking, which could not be faked.  Or IPsec could be used for

   the ping packets, making them secure.

* Security Summary

  So, in summary:  The only weakness in the Mandos system is from

  people who have:

  1. The power to come in and physically take your servers, *and*

  2. The cunning and patience to do it carefully, one at a time, and

     *quickly*, faking Mandos client/server responses for each one

     before the timeout.

  While there are some who may be threatened by people who have *both*

  these attributes, they do not, probably, constitute the majority.

  If you *do* face such opponents, you must figure that they could

  just as well open your servers and read the file system keys right

  off the memory by running wires to the memory bus.

  What this system is designed to protect against is *not* such

  determined, focused, and competent attacks, but against the early

  morning knock on your door and the sudden absence of all the servers

  in your server room.  Which it does nicely.