Types of Failure

So far you have learned what actions you may take to make yourself completely ready to recover fully from any type of security-related network failure. Although it is important to recover quickly from every failure, some types are more important than others, and therefore failures are handled differently. Network failures can also occur at different times:

• During the configuration/deployment phase As previously discussed, any addition, deletion, or change to the network (for example, a software upgrade, or adding a new device such as a PIX) should be accomplished during a scheduled maintenance window. If the change control is not implemented in due time, you must have a rollback capability to the working condition. If you encounter problems during configuration/deployment, most of the time this is considered to be a non-service affecting failure, and you have more time to analyze and troubleshoot the problem. For example, if you wanted to protect a segment of your network which is currently unprotected with a PIX firewall, and during PIX deployment you encounter problems with traffic, you could roll back to the previous setup (that is, leaving the network without PIX) until the root cause was known. With this type of failure, it is extremely important to collect the required information to get to the root cause of the problem.

• In the production network You could experience a failure in the production network. For example, traffic has been flowing without problems, when suddenly a user from the internal network cannot access the Internet across the PIX firewall. This type of failure is unexpected and requires in-depth knowledge of the product and a systematic approach to troubleshooting this type of failure. The next section covers these topics.