- The Life Cycle of an Incident
- Incident Reporting
- Keeping Current
- Writing Computer Security Advisories
- Summary
Incident Reporting
As previously noted in this chapter, every response team should have a report form that identifies the information it requires to investigate and track an incident. The information requested on the incident report form will vary from team to team, and should mirror specific data fields in the incident tracking database and/or trouble ticket system. Appendix A provides a sample form that may be used as a guide for an organization's own incident report form. The CERT CC report form is also available at http://www.cert.org and provides additional fields that may be considered.
The desired information should be clearly requested in the report form and allow little, if any, room for ambiguity. Pick lists or selection options can provide a tremendous advantage by eliminating confusion for both the reporting party and the incident handler. Pick lists are also extremely valuable in the incident database and trouble ticket systems. Without such lists, it's amazing how many different ways information may be defined and reported. Caution should be taken, however, to ensure that an “other” category is included for those cases that do not match any of the selections. The “other” selection should be accompanied by an area where specific comments may be made to expand upon the entry. Instructions for completing the report form should explain how to handle unclear areas or questions as they arise.
The report form should also ask for contact information for the reporting individual. In addition, the form should contain a place to indicate whether the e-mail address provided should not be used for communication regarding the incident. The ability to remain anonymous should also be considered, as some people may be reluctant to make a report if they have to provide their identity.
Feedback
Once the report form is completed and submitted to the incident response team, it should be acknowledged in some fashion. Lack of acknowledgment often leads to a feeling of helplessness and frustration on the part of the constituency. If the reporting party took the time to complete the report and submit it, then he or she should receive the satisfaction of a response indicating that someone was interested in the activity. The acknowledgment should include a unique incident or event tracking number and any pertinent information that needs to be passed on at that time. If additional information is needed, a request for it may accompany the response. At the same time, care should be taken to not disclose any information about the incident in the process of responding. The following provides a sample acknowledgment that may be used to respond to a report:
Thank you for your incident report dated June 15, 2002. We are analyzing the information reported to ascertain what may have caused the suspicious activity noted, and we are tracking this activity as event #0601-25.
We appreciate the inclusion of audit log excerpts with the report, but need to clarify the time zone for which these logs were recording activity. Please respond with the time zone used and reference the tracking number so we can update our records appropriately.
If you discover any additional information concerning this activity or have any questions, please feel free to contact us at 800-123-4567.
Regards,
(Person responding or team signature)
In some cases little, if anything, can be done regarding the activity reported (e.g., a scan of network addresses). An acknowledgment or response to the report is still warranted, however. Without it, the person completing the report may be reluctant to continue reporting in the future. In addition, despite the limited response capability that may be provided to that report, the information may still be vital to the monitoring of the organization's information security. The following is a sample acknowledgment to this type of report:
Thank you for your incident report dated December 1, 2002. We are tracking this event as incident #1201-01. Please reference this number should you discover any additional information concerning this incident.
While we may not be able to take action against this source based on the information you provided, your reports, along with those of other system administrators, will help us to understand the scope and frequency of these problems. Your information will be added to our database so we can correlate it with past activity.
If you discover additional information concerning this activity or have any questions, please feel free to contact us at 800-123-4567.
Regards,
(Person responding or team signature)
If activity is reported that turns out to not be an incident, then a response should be provided indicating that fact to the person who submitted the report. Depending on how the incident tracking numbers are assigned, the acknowledgment may include a tracking number. If applicable, the acknowledgment should be an awareness education opportunity in which feedback is provided describing why the activity was not considered to be an incident. If people are reporting nonincident activity, the education provided in the feedback can help to cut down on the number of unneeded reports.
Regardless of the type of acknowledgment sent, the response should be signed either by the incident handler who is responding to the report or with a team signature. Including an individual name can add to a sense of uniqueness to the process and detract from the feeling of a “canned” response, but it can also hinder follow-on communications regarding the incident. Specifically, if the person reporting discovers further information on the event and wishes to report it, that individual may believe that he or she can speak only with the person who signed the initial response. If that person is out for a few days or working a different shift, the unavailability can add to frustration for the reporting party and slow the communication. The person reporting should be encouraged to communicate with other members of the team as well, should they need to follow up on an incident.
Tracking Incidents
The number of incidents processed and responded to by every team will vary according to the size of the constituency. If an organization is large enough to establish its own incident response team, then the chances are very good that the team will need a database or mechanism to store and track incident data. Many larger teams have both a database and a trouble ticket system. The trouble ticket system helps with tracking the reports as they are received as well as already open items. Typically, an incident will be in one of three states from the incident response team's perspective:
-
Open: The incident has not been resolved and an action is required of the response team.
-
Waiting or pending: The incident has not been resolved, but the team is waiting for further information, an investigation to be conducted, or a response from another person or group.
-
Closed: No further action is required of the incident response team.
The trouble ticket system should annotate the current state of each incident and indicate what action is pending. It should also identify the flow of steps taken, so an incident handler responding for the first time to an incident can quickly see the history of action taken to date.
The trouble ticket system is the tool used to triage the requests and reports as they are received. Normally, the system would assist with identifying those incidents that are in the open or waiting state. Depending on the programs used for the system and the database, the data from the trouble ticket may be directly ported into the database. Likewise, if a closed incident is reopened, information previously entered into the database should be able to be recalled through the trouble ticket system if needed.
Working in combination, these systems can be extremely important tools for incident handling. Not only can they be used to track open incidents (ensuring that a report does not fall through the cracks) and store incident data, but they can also aid in the correlation of activity. The “correlation of activity” refers to the process of looking for patterns in activity that may be attributed to either the same source or the same type of attack. Running queries on the database can automate the steps taken to identify attacks stemming from the same source, targeting the same destination, using the same port or service, occurring during the same time period, involving the same “handle,” or following any other pattern. Without the database, this task can be tremendously time-consuming and the human eye may overlook important incident elements that are not readily apparent.
The database should be configured with enough storage to expand based on the number of records stored. As the incident response team becomes widely known and succeeds, the number of reports will increase exponentially, and the database must be able to support this growth. The statistics in Table 8-1 were taken from the CERT CC Web site and give the number of incidents processed over the years by that team. Although CERT CC can be considered as having the largest constituency and routinely handles more incidents than many other teams, the statistics should indicate how quickly the number of incidents can grow with time.
Another important use of the database is statistics generation. Statistics regarding the number of inquiries and incidents responded to can be used to help justify the hiring of additional people for the team or the purchase of additional resources. Statistics on parts of the organization experiencing the most successful attacks may indicate an area with training deficiencies or other problems that need to be resolved. Statistics indicating a decrease in successful incidents may be used to gauge the success of security programs or particular tools (e.g., antivirus software). Depending on the specific reporting needs of the organization, the data maintained in the database can be very useful for justifying the team's existence as well as pinpointing strengths and weaknesses in particular security programs.
Although an incident response team may use many tools, the database should be considered one of the (if not the) most important tool. Therefore, due consideration should be given to the type of data base purchased, the fields identified for tracking, and the hiring of a skilled database administrator for programming and maintaining the system. The database is not the area to cut costs or save dollars should funding problems emerge.
Table 8-1. Annual Number of Incidents Handled by the CERT CC
Year |
Number of Incidents Reported |
---|---|
1988 |
6 |
1989 |
132 |
1990 |
252 |
1991 |
406 |
1992 |
773 |
1993 |
1,334 |
1994 |
2,340 |
1995 |
2,412 |
1996 |
2,573 |
1997 |
2,134 |
1998 |
3,734 |
1999 |
9,859 |
2000 |
21,756 |
2001 |
52,658 |
2002 |
82,094 |
Total |
182,463 |