Researchers Say SNMP Open to Attacks
Security researchers have discovered several flaws in SNMP, a language used for communications with network devices, which they say could enable attackers to deny service to various devices or even take control of them entirely.
SNMP (simple network management protocol) is widely deployed on the Internet and in private networks and is used mainly to monitor and manage network devices. Recently, researchers at a Finnish university identified several vulnerabilities in SNMPv1 relating to the manner in which SNMP agents and managers handle messages.
The vulnerabilities affect products from dozens of vendors, including IBM Corp., Hewlett-Packard Co., Computer Associates International Inc., Cisco Systems Inc. and 3Com Corp. Most of the vendors have either released patches or are in the process of evaluating the problems as they affect their products.
There are multiple vulnerabilities in the way that many vendors products decode and process SNMP trap messages, which are sent from agents to managers, according to an advisory published by the CERT Coordination Center at Carnegie Mellon University in Pittsburgh.
The researchers at Oulu Universitys Secure Programming Group also found several problems with the way that agents handle and decode the request messages that managers send them.
The effects of exploiting one of these vulnerabilities will vary from product to product, CERT says, but in general, they can cause denials-of-service, service interruptions or an attacker gaining control of a vulnerable device.
CERT recommends that administrators disable SNMP on any machine that doesnt need it for normal operations, although the Finnish researchers discovered that some of the products they tested exhibited odd behavior or denial-of-service conditions even when SNMP was disabled.
The CERT advisory is available on the organizations site, www.cert.org.