So I understand the DSA/RSA authentication procedure and how it
provides authenication of a client from a servers perspective, but
what is there implemented in SSH to provide authentication of the
server from the clients perspective.? I know that on connection, the
server sends the client a key, identifying itself to the client. But
what prevents, a rogue client from grabbing the identification #, and
masking the servers IP, and setting up their own rogue server, with a
the same server id?
If this happend, the rogue server would most liekly not have he public
key of any client connecting to it, nor would any clients private keys
be comprimised b/c their keys are never transmitted. But the rogue
server, could present itself as the server u wanted. Is that left up
to the client to check, when they log in?

i.e: the client could place a file on the known good server, which
contains some text that only the client would know. then when they
logged in, they can authenticate the server, based on the existance
and content of the file, existing on the server.

>>>>> "NN" == Nate, Nano <nanonut@gmail.com> writes:

NN> So I understand the DSA/RSA authentication procedure and how it
NN> provides authenication of a client from a servers perspective, but
NN> what is there implemented in SSH to provide authentication of the
NN> server from the clients perspective.? I know that on connection,
NN> the server sends the client a key, identifying itself to the
NN> client. But what prevents, a rogue client from grabbing the
NN> identification #, and masking the servers IP, and setting up their
NN> own rogue server, with a the same server id?

The rogue server does not possess the corresponding private hostkey.

--
Richard Silverman
res@qoxp.net

>>>>> "NN" == Nate, Nano <nanonut@gmail.com> writes:

NN> So I understand the DSA/RSA authentication procedure and how it
NN> provides authenication of a client from a servers perspective, but
NN> what is there implemented in SSH to provide authentication of the
NN> server from the clients perspective.? I know that on connection,
NN> the server sends the client a key, identifying itself to the
NN> client. But what prevents, a rogue client from grabbing the
NN> identification #, and masking the servers IP, and setting up their
NN> own rogue server, with a the same server id?

The rogue server does not possess the corresponding private hostkey.

--
Richard Silverman
res@qoxp.net