Protecting Network Infrastructure at the Protocol Level

Contact:http://www.packetnexus.com

Protecting Network Infrastructure at the Protocol Level
 

V1.0. Curt Wilson, Netw3.com Consulting. 12/15/00

 

This paper is posted in the Netw3.com security reading room and can be
found along with other security documents at
http://www.netw3.com/documents.html

 

Scope of paper
 

This paper will briefly discuss attacks and attack prevention methods
for network infrastructure protocols. Particular focus will be given to
router and routing protocol vulnerabilities such as Routing Information
Protocol (RIP), Border Gateway Protocol (BGP), Open Shortest Path First
(OSPF), and others.

 

Routers perform a critical function for each network and if a router is
compromised or a route is successfully spoofed, network integrity can be
seriously damaged especially if hosts are not using encrypted
communications channels. The potential for data manipulation through
man-in-the-middle attacks, denial of service, data loss, disruption of
network integrity, and packet sniffing is great. Security mechanisms are
often available, but are commonly not used because attacks on routing
protocols have been rare. Due to the lack of hard data on actual
incidents, some approaches outlined in this paper will be theoretical in
nature.

 

Routing is a huge and complex topic; therefore this document will be
updated and corrected as I continue my research. Note that I am not a
routing engineer and would be glad to accept corrections to any
information contained herein.

 

Commonly known router security issues
 

Various types of routers have well-known security issues. A collection
of some of the commonly known vulnerabilities for network infrastructure
equipment vendors such as Cisco, Livingston, Bay and others, can be
found at
http://www.antionline.com/cgi-bin/anticode/anticode.pl?dir=router-exploi
ts. Most of these vulnerabilities are non routing-protocol level attacks
that rely on misconfiguration, bugs in IP packet handling, SNMP
insecurities such as default community name strings, weak password or
weak password encryption, DOS conditions due to bad IP/UDP packets, etc.
These types of attacks are commonly known, and a standard NIDS should be
able to be programmed to detect these, at least on an IP based network.
IDS are still in the emerging stages as far as non-TCP/IP based routing
protocols are concerned. Any of these types of attacks can weaken a
network infrastructure and could be used in combination with
higher-level protocol-based attacks.

 

Proper configuration management can resolve many of these common
vulnerabilities. This would involve standard procedures such as not
using SNMP (or choosing strong passwords/encryption), keeping up to date
with vendor patches, proper use of access lists, ingress/egress
filtering, firewalls, encrypted management channels and passwords, route
filtering, and use of MD5 authentication. However, to understand and
implement these security procedures, network engineers must be given the
time and training to understand the security implications of their work

 

 

Recent Developments in Infrastructure Defense
 

A recent development in network defense comes in an IDS called JiNao,
which can be found at http://www.anr.mcnc.org/projects/JiNao/JiNao.html.
JiNao is funded by DARPA and is currently in development as a joint
research project between MCNC and North Carolina State University.