Re: Problem with ascend pipeline routers.

From: James Bass (jamesbat_private)
Date: Fri May 29 1998 - 10:33:59 PDT

  • Next message: Richard Braakman: "Re: ircii-pana (BitchX) 74p4 overflow - exploit/fix"

    You can always set up secure-access on it (if you want to waste a few
    bucks), or just set up a few filters so that only certain boxes (or only
    the LAN) have access to telnet to the box.
    
    There is a FAQ that addresses that issue at:
    http://www.ascend.com/faqs/400/122.html
    
    
    
    > -----Original Message-----
    > From: Joe Shaw [SMTP:jshawat_private]
    > Sent: Thursday, May 28, 1998 3:54 PM
    > To:   BUGTRAQat_private
    > Subject:      Re: Problem with ascend pipeline routers.
    >
    > On Wed, 27 May 1998, Eric Thacker wrote:
    >
    > > Date: Tue, 26 May 1998 14:19:30 -0700
    > > From: support <supportat_private>
    > > To: ericat_private
    > > Subject: RE: Ticket #238563
    > >
    > > Eric:
    > >
    > > The pipeline has no way of telling what is a legit telnet and what
    > is
    > > not and because the box is meant to be accessed this way (both
    > locally
    > > and remotely), a firewall is the best way to restrict telnet access.
    > >
    > > --
    > > Ascend Communications, Inc          Service & Support
    > > supportat_private
    > > http://www.ascend.com/service
    >
    > I've forwarded this to the ascend users group (ascend-usersat_private)
    > so
    > we can get Kevin Smith's (kevinat_private) and Matt Holdrege's
    > (mattat_private) opinion on this.
    >
    > I have my own opinions on the problem.  The Pipeline family has always
    > been a very basic, barebones line of routers with a somewhat limited
    > set
    > of functions.  They'll do NAT, RIP, etc. but all of them only allow
    > you
    > two telnet sessions into the router and only one diagnostics session.
    > This was done to limit the resources that would be consumed so ram/cpu
    > wouldn't be burderend with tons of telnet sessions and/or diagnostic
    > sessions which can kill it's bigger brother the MAX, just like doing
    > debug all on a cisco will hose it.
    >
    > But regardless, even if there was an idle timer on the authentication
    > mechanism, there are still a few problems.  One problem is that even
    > with a login timeout, you're going to have two telnet sessions max,
    > which
    > is a limit placed most likely by the resources mentioned earlier.
    > So,
    > you can just keep opening telnet sessions as soon as the others die
    > and
    > see if you can keep telnet access locked up.  Also, as of the 5.0A
    > versions of code (and possibly earlier) the ascend doesn't log telnet
    > connections to the syslog facility.  My maxen don't do it, and the
    > pipeline logs I've looked at don't do it.  They do log incoming calls,
    > call up, call down, and various other events, but not telnet access.
    > So, tracking down who's doing this would involve a sniffer, which
    > might
    > not be a readily available tool to most people with pipelines as
    > routers.
    >
    > You're right though, the best way to handle the problem is by
    > firewalling
    > or filtering out telnet access.  Guess this is a good selling point
    > for
    > their Secure Access Firewall option.
    >
    > Also, I was able to knock out 3 of my Ascend Maxen by repetedly
    > telneting
    > to them.  Software version was 5.0Ap51, file ti.m40.  On a heavily
    > loaded
    > max, I opened two successful simultaneous telnets and it died on the
    > third, while another max died at four.  I'm going to investigate
    > further,
    > and I have no info on the 6.0.X versions of max code yet.
    >
    > Joe Shaw - jshawat_private
    > NetAdmin - Insync Internet Services
    



    This archive was generated by hypermail 2b30 : Fri Apr 13 2001 - 13:55:22 PDT