Securing SNMP

SNMP - Simple Network Management Protocol

A popular protocol that is widely utilized for network management. SNMP (RFC 1157) can be very useful for network and systems administrators as it can provide real-time alerts and remote administration options (such as restarting a device over a network). On paper this protocol sounds like the perfect compliment to the IT professional, however it can actually cause more heartache then good if not planned and implemented properly. While reading this article please try not to get lost in the lingo and keep in mind that the basic point of SNMP is for communication between 2 machines.

How does SNMP work?

The Simple Network Management Protocol is part of the IP suite and resides on the applicaton layer of the OSI model. There are other aspects that define the protocol, including a database schema and data objects. The main components to any SNMP system are:

1. Monitored/managed devices: The device that we want to manage
   
2. SNMP Agent: Acts as the communicator between the NMS and Managed Device
   
3. MIB (Management Information Bases): Decides which data to collect on the managed device
   
4. Network management systems (NMSs): Collects the information about the managed device and can allow for remote management.
   

Examining the following diagram you will see how the components interact with each other:


Operation Modes

There are 2 basic modes in which information can be collected. In the more generic form of SNMP you may have your NMS poll each managed device for information. This is fine unless you are managing a lot of devices, as response time would slow. To address this the other basic mode is to have the SNMP agent on the managed devices send the information to the NMS. This is also referred to as an SNMP Trap, because the agent is trapping the information and then sending a Trap (or INFORM) signal to the NMS. The NMS can then be configured to accept trap messages.

Security Considerations

It is important to note that SNMP is considered to be insecure and should not be used if not needed. There are ways to secure it, but first it is important to understand the vulnerabilities.

1. There are 3 versions of SNMP, version 1, 2, & 3. The current and most highly recommended implementation is version 3.
   
2. Versions 1 & 2 are passed over a network in clear text and only version 3 includes encryption.
   
3. Windows 2003 utilizes only Versions 1 & 2.
4. All versions of SNMP are vulnerable to brute force cracking

Keeping these flaws in mind, there are ways to address these issues.

• If possible, do not use SNMP
  
• Use the most recent version of SNMP (v3) if supported
• If v1 or v2 is used, you may consider encrypting network traffic with ipsec
• Consider utilizing an administrative VLAN segregated away from a VLAN with workstations
  
• Do not use default community strings (you wouldn't use a generic password that everyone knows; this is the same concept)
• Use a community string that is not easily guessed or cracked (utilize numbers and special characters and absolutely no dictionary words)
• Restrict access to the agents to only the SNMP Management IP(s)
• If read/write access if not required then only allow Monitor access on the agent and, if hardware permits, on the switch.
  
• In conjunction with the previous bullet, if write access is not required and hardware permits then Disable the SET command, which allows the SNMP agent to be configured.
• If not needed, some hardware will allow you to disable the option for rebooting the device over the network.
• Lastly, consider the use of an IDS (Intrusion Detection System), as it can alert you of any malformed packets or traffic that is unusual or out of place.

Conclusion

SNMP can be a great tool for network management, assuming the need is there. There are other forms of network management which may be more suited to your needs, however if done correctly and securily SNMP is a viable option.

References

Wikipedia

Cisco

Add Date and Time Stamps to Your Batch Files

Ever want to setup a windows batch file to output a date or time without using vbscript? My specific reason for researching this function was to dynamically generate a new log file each time my script was run. The following code will set the date and time into variables names %cdate% and %ctime% and then generate a text file named backup-DATE-TIME.txt (obviously replacing DATE and TIME with the current date and time).

@echo off
REM This section sets the date variables
FOR /F "tokens=1-4 delims=/- " %%a in ('date/T') do set CDATE=%%b%%c%%d
REM The following formats the date variable to MMDDYY
SET CDATE=%CDATE:~0,4%%CDATE:~-2%
FOR /F "tokens=1-4 delims=:., " %%a in ("%TIME%") do set CTIME=%%a%%b%%c
echo.
echo. > c:\logs\backup-%cdate%-%ctime%.txt
echo.

This script can be inserted into almost any script fairly easy and you can utilize the date and time variables anywhere in the script.

References:
Batchfiles FAQ

Setting up OpenSUSE 10.3 to use Windows Authentication

The following utilizes Yast as the setup method. You may also setup manually, installing packages, editing the conf files and running the commands... or enter a few options and have Yast do the work. There are benefits to setting up domain membership manually, but this howto is a quick join method.

1. Launch Yast
2. Choose Windows Authentication
3. Allow the additional packages to install (if necessary)
4. Type your Domain in the following format:
1. testdomain.com
5. Check Create Home directory on login and offline authentication (both explanatory)
6. If you would like users to share their home directories (on this machine) then select the options.
7. Setup NTP
8. Click next and it prompt you for username and password (if all goes well)
9. It should now confirm that you've joined the domain.
10. Fin!
    

Refernences:
Novell Documentation