It is important that an incident response plan is formulated, supported throughout the organization, and is regularly tested. A good incident response plan can minimize not only the affects of the actual security breach, but it may also reduce the negative publicity.
From a security team perspective, it does not matter whether a breach occurs (as such occurrences are an eventual part of doing business using an untrusted carrier network, such as the Internet), but rather, when a breach occurs. Do not think of a system as weak and vulnerable; it is important to realize that given enough time and resources, someone can break into even the most security-hardened system or network. You do not need to look any further than the Security Focus website, http://www.securityfocus.com/, for updated and detailed information concerning recent security breaches and vulnerabilities, such as the frequent defacement of corporate webpages or the 2002 attacks on the root DNS nameservers.
The positive aspect of realizing the inevitability of a system breach is that it allows the security team to develop a course of action that minimizes any potential damage. Combining a course of action with expertise allows the team to respond to adverse conditions in a formal and responsive manner.
The incident response plan itself can be separated into four phases:
Immediate action to stop or minimize the incident
Investigation of the incident
Restoration of affected resources
Reporting the incident to the proper channels
An incident response must be decisive and executed quickly. Because there is little room for error, it is critical that practice emergencies are staged and response times measured. This way it is possible to develop a methodology that fosters speed and accuracy, minimizing the impact of resource unavailability and potential damage in the event of an actual system compromise.
An incident response plan has a number of requirements, including:
A team of in-house experts (a Computer Emergency Response Team)
A legally reviewed and approved strategy
Financial support from the company
Executive/upper management support
A feasible and tested action plan
Physical resources, such as redundant storage, standby systems, and backup services
The Computer Emergency Response Team (CERT) is a group of in-house experts who are prepared to act quickly in the event of a catastrophic computer event. Finding the core competencies for a CERT can be a challenge. The concept of appropriate personnel goes beyond technical expertise and includes logistics such as location, availability, and desire to put the organization ahead of ones personal life when an emergency occurs. An emergency is never a planned event; it can happen at any moment and all CERT members must accept the responsibility that is required of them to respond to an emergency at any hour.
CERT teams typically include system and network administrators as well as information security experts. System administrators provide the knowledge and expertise of system resources, including data backups, backup hardware available for use, and more. Network administrators provide their knowledge of network protocols and the ability to re-route network traffic dynamically. Information security personnel are useful for thoroughly tracking and tracing security issues as well as performing a post-mortem (after the attack) analysis of compromised systems.
Although it may not always be feasible, there should be personnel redundancy within a CERT. If depth in core areas is not applicable to an organization, then cross-training should be implemented wherever possible. Note, if only one person owns the key to data safety and integrity, then the entire enterprise becomes helpless in that one person's absence.
Some important aspects of an incident response to consider include legal ramifications. Security plans should be developed with members of legal staff or some form of general counsel. Just as every company should have their own corporate security policy, every company should have its own way of handling incidents from a legal perspective. Local, state, and federal regulatory issues are beyond the scope of this document, but are mentioned because the methodology for performing a post-mortem analysis, at least in part, is dictated by (or in conjunction with) legal counsel. General counsel can alert technical staff of the legal ramifications of security breaches; the hazards of leaking a client's personal, medical, or financial records; and the importance of restoring service in mission-critical environments such as hospitals and banks.