OSEC

Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com
 
From: Attonbitus Deus (ThorHAMMEROFGOD.COM)
Date: Fri Apr 06 2001 - 12:46:14 CDT

  • Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

    This is for the inbound packet, so think of it from the packet's
    perspective:

    You need to allow any udp packet with source port > 1024 with any source
    address back in if the destination address is in your subnet and the
    destination udp packet is 53. The client queries 53 from >1024- that guy
    has to come back in...

    I applied this filter and it works perfectly.

    HTH
    ---------------------------------
    Attonbitus Deus
    ThorHammerofGod.Com

    ----- Original Message -----
    From: "Phil Cox" <Phil.CoxSYSTEMEXPERTS.COM>
    To: <FOCUS-MSSECURITYFOCUS.COM>
    Sent: Thursday, April 05, 2001 4:10 PM
    Subject: Re: Win2K port filtering, DNS

    > > I'm trying to add Win2K port filtering (not via IPSec) to a Win2K
    > > IIS Server, and it breaks dns as a client. I've allowed both 53 UDP and
    > TCP.
    >
    > Remember that the client will use a "random" numbered high port (>1024
    > typically) as the source port, and the destination is 53. So adding these
    > filters will not do anything from the clients perspective.
    >
    > > Based on Philip Cox's 'Hardening Windows 2000' paper
    > > (http://www.systemexperts.com/tutors/HardenW2K101.pdf)
    > > I don't think I should even have to explicitly open port 53
    > > to use dns as a client.
    >
    > You don't, TCP/IP port filtering is only for inbound connections.
    >
    > > In any case, with port filtering on, and 53 (tcp and udp) open,
    > > dns times out. With port filtering off, dns works fine.
    >
    > What happens if you allow all UDP ports? I am wondering if it is the case
    > that by blocking UDP packets, it block even the return packets (i.e.,
    > srcport = 53, destport >1024), I thought there was a bit of statefulness
    in
    > it, but I'll have to check again.
    >
    > Nope, ... I get the same results. I'll have to investigate this, as I
    don't
    > have the answer off the top of my head ;(
    >
    > Phil