1:

If you need to use Telnet in-band to manage routers, are there any other security techniques that can limit the damage done if an attacker is able to capture the packets of your session?

A1:

First, using OTP and user-specific login will prevent the attacker from learning a usable password that could be used to log in to the device. In addition, using an "enable secret" password instead of a basic enable password will prevent the attacker from learning the password as it passes by in a configuration file viewed remotely.

2:

If you were operating an ultrasecure environment and were concerned about the possibility of PVLAN bypass in the OOB design, is there anything else you could do to add to the security?

A2:

You could break up the OOB network into zones in much the same way as the rest of your network. Instead of all devices coming into one PVLAN L2 network, you could have your Internet edge devices on one and your campus network on another. Each network could terminate on a discrete interface on the management firewall.

3:

Is it acceptable to run read-write SNMPv1 on your main corporate firewall?

A3:

If you answered "it depends," congratulations! All management functions are beholden to the same security concepts in the rest of the book. Business requirements, policies, and risk analysis are all factors, as are the security capabilities of the device. If, for example, you are tunneling the SNMP traffic over IPsec, the risk is reduced. All this said, "no" tends to be the answer to this question for most networks.

4:

Will NETOPS folks ever need to manage aspects of a device with principally security-related functions?

A4:

Sure; the corporate firewall is a fine example. The NETOPS team is often charged with performance management. Monitoring the performance of the firewall in passing all inbound and outbound Internet traffic could be considered a critical component of the overall edge network. This level of access might be granted with SNMP read-only capabilities on the firewall.

5:

With the diversity of data that a Syslog server can receive, how is it possible for the device to deal with event ordering?

A5:

Although it is possible to time stamp Syslog data when it arrives at the Syslog server, it is best to time stamp data using Network Time Protocol (NTP) before it leaves the managed device. NTP is discussed in Chapter 5, "Device Hardening," and should be implemented on almost all devices that support it. Additionally, Syslog supports different "facilities" to log different types of messages to different files. Refer to the documentation of your Syslog tool for more details.

Категории