SNMP Vulnerabilities: Read Community String Guessable

Impact

A read community string for one of your systems can be easily guessed. The full impact will depend largely on exactly what type of device this system is. In general, anyone guessing this read community string can obtain an awful lot of information regarding the device in question, and possibly about the network(s) it is on. You need to decide whether this is a security concern or not.

Background

The Simple Network Management Protocol (SNMP) is a widespread protocol allowing network administrators to obtain information on and even configure various network devices remotely. It is very common on all but the most basic networking hardware (hubs, switches, routers, etc), and many other networked devices (networked printers, terminal servers, etc). Many workstations/PCs also have SNMP clients running on them as well, and most network management packages (commercial and non-commercial) make extensive use of SNMP for information gathering.

Most devices that provide SNMP allow enormous amounts of data to be accessed over it. The exact information available depends on the type of device, its manufacturer and model, but generally include details of the hardware and OS type, information on the various network interfaces, statistics on the various network protocols, and general and vendor-specific details about what the device does and is doing. The volume of data available is generally too much to be useful to a systems administrator without some management code to sort through it. The security risks of allowing a potential intruder access to this information depends largely on what type of device it is, but realize that if the data is known to the device, it is probably accessible via SNMP.

Many devices allow themselves to be configured remotely via SNMP as well. Devices which do so generally can be completely configured in such a manner. This can definitely be of use to systems administrators, but also is an obvious security concern.

Despite its popularity, SNMP v1 and v2 have rather basic access control, using passwords called community strings. Most devices are set up with two community strings, a (Read) community for viewing information and a Set or Write community for changing configurations. Many devices come out of the box with SNMP operational and a read community string of "public". Write access often has to be turned on manually, but not always. Needless to say, care should be taken with both settings.

The Problem/Resolution

If you were notified of this vulnerability, a read or write community string was able to be guessed for a system you scanned. This is currently done with a simple, brute force algorithm, repeated trying a few guesses. In order to guess write community strings, it actually attempts to change the sysLocation oid (and then changes it back if succeeded). If it guessed your community string, you should consider changing it.

Some SNMP clients will allow you to restrict which hosts can send some or all write SNMP commands from, and possibly which hosts can get information as well. It is recommended that you configure such if available.

Related CVE entries:
CVE 2001-0514 Atmel 802.11b VNET-B Access Point

Where can I read more about this?

For more information on SNMP, see Cisco's SNMP Reference.