new tcp scan method

antirez (antirez@SECLAB.COM)
Fri, 18 Dec 1998 07:47:57 +0100

  Hi,

        I have uncovered a new tcp port scan method.
        Instead all others it allows you to scan using spoofed
        packets, so scanned hosts can't see your real address.
        In order to perform this i use three well known tcp/ip
        implementation peculiarities of most OS:

          (1) * hosts reply SYN|ACK to SYN if tcp target port is open,
            reply RST|ACK if tcp target port is closed.

          (2) * You can know the number of packets that hosts are sending
            using id ip header field. See my previous posting 'about the ip
            header' in this ml.

          (3) * hosts reply RST to SYN|ACK, reply nothing to RST.


        The Players:

          host A - evil host, the attacker.
          host B - silent host.
          host C - victim host.

        A is your host.
        B is a particular host: It must not send any packets while
          you are scanning C. There are a lot of 'zero traffic' hosts
          in internet, especially in the night :)
        C is the victim, it must be vulnerable to SYN scan.

        I've called this scan method 'dumb host scan' in honour of host
        B characteristics.


        How it works:

        Host A monitors number of outgoing packets from B using id iphdr.
        You can do this simply using hping:

#hping B -r
HPING B (eth0 xxx.yyy.zzz.jjj): no flags are set, 40 data bytes
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=0 ttl=64 id=41660 win=0 time=1.2 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=1 ttl=64 id=+1 win=0 time=75 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=2 ttl=64 id=+1 win=0 time=91 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=3 ttl=64 id=+1 win=0 time=90 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=4 ttl=64 id=+1 win=0 time=91 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=5 ttl=64 id=+1 win=0 time=87 ms
-cut-
..
.

        As you can see, id increases are always 1. So this host have the
        characteristics that host B should to own.

        Now host A sends SYN to port X of C spoofing from B.
        (using hping => 0.67 is very easy, http://www.kyuzz.org/antirez)
        if port X of C is open, host C will send SYN|ACK to B (yes,
        host C don't know that the real sender is A). In this
        case host B replies to SYN|ACK with a RST.
        If we send to host C a few of SYN it will reply to B with a few
        of SYN|ACK, so B will reply to C a few of RST... so
        we'll see that host B is sending packets!

.
..
-cut-
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=17 ttl=64 id=+1 win=0 time=96 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=18 ttl=64 id=+1 win=0 time=80 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=19 ttl=64 id=+2 win=0 time=83 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=20 ttl=64 id=+3 win=0 time=94 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=21 ttl=64 id=+1 win=0 time=92 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=22 ttl=64 id=+2 win=0 time=82 ms
-cut-
..
.

        The port is open!

        Instead, if port X of C is closed sending to C a few
        of SYN spoofed from B, it will reply with RST to B, and
        B will not reply (see 3). So we'll see that host B is not sending
        any packet:

.
..
-cut-
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=52 ttl=64 id=+1 win=0 time=85 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=53 ttl=64 id=+1 win=0 time=83 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=54 ttl=64 id=+1 win=0 time=93 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=55 ttl=64 id=+1 win=0 time=74 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=56 ttl=64 id=+1 win=0 time=95 ms
60 bytes from xxx.yyy.zzz.jjj: flags=RA seq=57 ttl=64 id=+1 win=0 time=81 ms
-cut-
..
.

        The port is closed.

        All this can appear complicated to perform, but using two sessions
        of hping on Linux virtual consoles or under X makes it more simple.
        First session listen host B: hping B -r
        Second session send spoofed SYN: hping C -a B -S

        Sorry if my english is not so clear.
        However this posting is not adequate to describe exaustively
        this scan method, so i'll write a paper on this topic, specially
        about how to implement this in a port scanner (i.e. nmap), and
        about players characteristics and OS used.

happy new year,
antirez

--
Salvatore Sanfilippo
Intesis SECURITY LAB            Phone: +39-02-671563.1
Via Settembrini, 35             Fax: +39-02-66981953
I-20124 Milano  ITALY           Email: antirez@seclab.com