Re: [Qemu-devel] [PATCH 6/6] vnc: track & limit connections

[Daniel P. Berrange]

On Fri, Oct 17, 2014 at 02:34:07PM +0800, Gonglei wrote:
> On 2014/10/16 18:46, Gerd Hoffmann wrote:
> 
> >   Hi,
> > 
> >>> I try to prevent that by dropping the *oldest* connection, so you have a
> >>> chance to connect even if a unprivileged attacker tries to use up all
> >>> connection slots.
> >>
> >> Lets say the limit is 5. The bad guy has 5 open idle connections.
> >> The good guy opens a new one and pushes off one of the bad guy's
> >> connections. Fine so far. The bad guy though can simply open 5 more
> >> connections and he'll push the good guy's connection off again.
> > 
> > Correct.  It can't fully prevent the attack, but makes it harder to pull
> > off.  Just having $limit idle connects isn't enough any more, the bad
> > guy has to constantly bomb qemu with vnc connect requests, hoping this
> > kicks out the good guy before it managed to authenticate.  The chances
> > for the good guy are a bit better and it is also more likely that the
> > attack sets off alarms in network monitoring.
> > 
> 
> Hi,
> 
> Happy that I don't miss this patch series and conversation.
> I have a approach to prevent the brute force attack (which
> had been tested in my team). Firstly, we must set password for
> vnc server for security. Secondly, the DoS may bomb qemu
> with vnc connect requests, trying to decrypt password at present.

Note that VNC passwords offer no meaningful level of security.

If you want security for VNC you must *always* use the TLS extension
or the SASL extension, or both. These offer proven cryptographically
strong authentication protocols.

> If we set the max trying times, and then
> There are some concepts:
>  - INTERVAL_TIME: a time window that user can connnet vnc server
>  - REJECT_TIME: the time of reject any connection
>  - MAX_TRY_TIMES: the times that user can connect vnc server in INTERVAL_TIME,
>    if attach the MAX_TRY_TIMES, the server will lock, any user can not 
> connect again
>    before REJECT_TIME attached. The old connected client will not be 
> influenced.

How are you defining "user" in this description. Do you mean "Source IP address"
here ? Or any client connection ? Or something else ?

Regards,
Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://autobuild.org       -o-         http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org       -o-       http://live.gnome.org/gtk-vnc :|