[Date Prev][Date Next]
[Chronological]
[Thread]
[Top]
Details on TCP sequence numbers (RE: C API: minor comments)
- To: "Paul Leach (Exchange)" <paulle@Exchange.Microsoft.com>, "'Kurt D. Zeilenga'" <kurt@boolean.net>, Mark Wahl <M.Wahl@INNOSOFT.COM>
- Subject: Details on TCP sequence numbers (RE: C API: minor comments)
- From: Harald Tveit Alvestrand <Harald@Alvestrand.no>
- Date: Tue, 16 Nov 1999 08:08:32 -0500
- Cc: "Paul Leach (Exchange)" <paulle@Exchange.Microsoft.com>, mcs@netscape.com, howes@yahoo.com, "Andy Herron (Exchange)" <andyhe@Exchange.Microsoft.com>, "Anoop Anantha (Exchange)" <anoopa@Exchange.Microsoft.com>, kurt@OpenLDAP.Org, ietf-ldapext@netscape.com
- In-reply-to: <19398D273324D3118A2B0008C7E9A569051BEAAE@SIT.platinum.corp.microsoft.com>
- Resent-date: Mon, 15 Nov 1999 23:24:25 -0800 (PST)
- Resent-from: ietf-ldapext@netscape.com
- Resent-message-id: <"htJ-YB.A.AqD.UaQM4"@glacier>
- Resent-sender: ietf-ldapext-request@netscape.com
At 18:07 15.11.99 -0800, Paul Leach (Exchange) wrote:
> If this is a security risk, I suggest adding a security consideration to
> both RFC2251 and this draft stating the applications concerned
> about spoofing should utilize a secure transport.
Use of random initial sequence numbers has been known for a long time to
improve the security of TCP against session hijacking attacks, without
requiring the use of a secure transport, which will be unavailable for
anonymous connections. I see no reason not to recommend it here.
The particular attack being guarded against by TCP random initial sequence
numbers is described in RFC 1948; it lets an attacker set up a connection
from an IP address other than his own even when he is not able to listen to
the packets coming back, so as to be able to (for instance) get a command
executed over the Berkeley Remote Shell interface.
It only works for impersonating a client, not impersonating a server.
For the attack to be successful, it depends on:
- Trust being extended on the basis of source IP address
- Initial sequence numbers being badly chosen
It does not seem particularly relevant to LDAP, since:
- LDAP does not rely on IP addresses for authentication (shouldn't, anyway!)
- Both strong authentication and secure transport for anonymous clients are
available
in LDAP (you can use TLS with anon clients just fine - see 99% of the
HTTPS out there)
- An attack that depends on faking message IDs without the ability to
listen in on the
channel requires TCP transport hijacking, not just the RFC 1948 attack,
since one
needs a legitimate user to get past the authentication phase first.
This is a lot tougher than the RFC 1948 attack; I'd say it's impossible
without being
able to listen to at least one direction of traffic, and then LDAP
message IDs are
available anyway.
Security: the devil is in the details....
Harald A
--
Harald Tveit Alvestrand, Maxware, Norway
Harald.Alvestrand@maxware.no