Security Incidents mailing list archives

Re: Unidentified Trojan?


From: EPerrin () METROLAND COM (Elliot Perrin)
Date: Thu, 18 May 2000 17:11:15 -0400


I checked RFC 1700 and it says that port 542 is unassigned, however
my FreeBSD box /etc/services file says that port 542 is used for
ncp, which i have not seen before and am unfamiliar with.

Anyone know what ncp is, all I could find related to ncp as an acronymn was
network control protocol which was on a page related to PPP.

-----Original Message-----
From: Richard Ginski [mailto:rginski () CO PINELLAS FL US]
Sent: May 18, 2000 11:55 AM
To: INCIDENTS () SECURITYFOCUS COM
Subject: Unidentified Trojan?


We have been monitoring some strange activity regarding a
possible trojan on some of our systems. Unfortunately, this
explanation has to be long, in order to paint the whole picture:

1) We first noticed that there was a problem when we noticed
that two of our INTERNAL DNS servers appeared to be affected
by DNS cache poisoning. It was stumbled on accidentally when
someone entered a typo in a URL (ommitting a ":" when
specifying a port number to one of our intranet sites) and
was re-directed to a porn site: 216.65.124.73
(internalmachine.domain&portnumber). We figured it was cache
poisoning because I could not fathom that the DNS servers
would "learn" a host address for which there are no root servers for.

2) I checked our firewall logs as to who may have also
(involuntarily) been connecting to this IP address
(216.65.124.73) and found over 25 machines trying to connect
to this site using different port numbers (not HTTP). First,
the machines used ping (we don't allow outbound ping), then
used various ports. Finally, the machines just tried to
connect to this site using CIFS.

It appears that once the machines are turned on, the "trojan"
activity would begin. We tried to narrow down what could be
causing this (activity went on for two days) then the
activity ceased. Anti-virus software has always been
installed on these machines (Inoculan) and we manually
scanned one of the machines just to make sure the real time
scanner did not miss anything. Nothing was found. The dates
for which this occurred were 4/26 and 4/27. During those two
days were able to restart/login to these machines and watch
the activity a sniffer as we tried to determine the culprit.

3) We felt we had taken a number of precautions to prevent
any further damage, including, notification when any more
attempts were made to connect to the IP address 216.65.124.73.

4) Well, it started happening again on Tuesday of this week
(5/16) and continued till yesterday (5/17). It appears that
now the "destination port of choice" is TCP port 524 to the
same IP address, for which I can not identify for any type of
service. Approximately, 25 machines (different machines than
the machines before, on different network segments)  were
affected. Unlike before, we could not reboot/login to these
machines and cause them to make additional connection
attempts which seemed to stimulate the activity before.

5) Today (5/18), no connect attempts were made to
216.65.124.73. However, doing a search on destination port
524 revealed that machines are now trying to connect to some
of our HTTP servers in our DMZ.

All of the machines affected are Windows based (95/98 and NT).

To the best of our knowledge, all attempts to connect to this
outside address have failed due to our firewall.

Has anyone had any experience with this behavior? Can anyone
identify TCP port 524?

Any input would be greatly appreciated!



Current thread: