|
Posted by on October 4, 2005, 10:17 pm
If you were Registered and logged in, you could reply and use other advanced thread options
>In this case, you are setting up a man in the middle - your SSL
>communications are secured only between your PC and LogMeIn in each
>instance, not between the two PCs. The LogMeIn server is seeing a 'clear'
>version of your network traffic, because it is the end-point of both SSL
>sessions.
Actually, this is not true in this case. (Granted, Alun seems to know
his SSL, but that does not mean he's 100% correct.) LogMeIn runs with a
modified SSL stack on the host side - the traffic is end-to-end
encrypted between the browser and the computer being accessed. You can
verify this with any packet logger - simply capture a remote session on
both the browser and the host side. The first few kilobytes (SSL
session negotiation) will definitely differ, but then the data stream
will become identical after a short while. This would not be possible
if the LogMeIn service were decrypting and re-encrypting the traffic.
The modified SSL stack allows the service to mediate the initial
connection and then lets the server (host) take over with its own
re-negotiated keys.
LogMeIn also includes a peer-to-peer layer based on UDP NAT traversal
that kicks in a few seconds into the connection. When this happens the
(still SSL-encrypted) traffic simply avoids the LogMeIn datacenters and
flows directly between the ActiveX/Mozilla plugin in the browser and
the computer being accessed.
>Of course, your most likely protection in this case is that the LogMeIn
>service has no interest in doing so, and would stand to lose more if found
>doing so than they could possibly benefit. If that's not the case, then you
>need to look at strongly vetting the LogMeIn service and its staff.
I guess this also applies to any sort of Internet-enabled software you
decide to install on your computer.
|