Security event manager

From Wikipedia, the free encyclopedia
Jump to: navigation, search

A security event manager (SEM) (acronyms SIEM and SIM) is a computerized tool used on enterprise data networks to centralize the storage and interpretation of logs, or events, generated by other software running on the network.[1][2][3]

SEMs are a relatively new idea, pioneered in 1999 by a small company called E-Security, and in 2010 are still evolving rapidly. Often confused with security information managers (SIMs) and security information and event managers (SIEMs). The key feature of a Security Event Management tool is the ability to analyse the collected logs to highlight events or behaviors of interest, for example an Administrator or Super User logon, outside of normal business hours. An adjacent, but somewhat different market also exists for Log Management; although these two fields are closely related, log management typically focuses on collection and storage of data whereas SEM focuses on data analysis. Some vendors specialize in one market or the other and some do both, or have complementary products.

Many systems and applications which run on a computer network generate events which are kept in event logs. These logs are essentially lists of activities that occurred, with records of new events being appended to the end of the logs as they occur. Protocols, such as Syslog and SNMP, can be used to transport these events, as they occur, to logging software that is not on the same host on which the events are generated. The better SEMs provide a flexible array of supported communication protocols to allow for the broadest range of event collection.

It is beneficial to send all events to a centralized SEM system for the following reasons:

  • Access to all logs can be provided through a consistent central interface
  • The SEM can provide secure, forensically sound storage and archival of event logs (this is also a classic log management function)
  • Powerful reporting tools can be run on the SEM to mine the logs for useful information
  • Events can be parsed as they hit the SEM for significance, and alerts and notifications can be immediately sent out to interested parties as warranted
  • Related events which occur on multiple systems can be detected which would be impossible to detect if each system had a separate log
  • Events which are sent from a system to a SEM remain on the SEM even if the sending system fails or the logs on it are accidentally or intentionally erased

In addition to collecting and storing data, SEMs distinguish themselves from simpler log management tools by providing a deeper level of event analysis. This may include attaching contextual information, such as host information (value, owner, location, etc.), identity information (user info related to accounts referenced in the event like first/last name, workforce ID, manager's name, etc.), and so forth. This contextual information can be leveraged to provide better correlation and reporting capabilities and is often referred to as Meta-data.

SEMs can also integrate with external remediation, ticketing, and workflow tools to assist with the process of incident resolution. The better SEMs will provide a flexible, extensible set of integration capabilities to ensure that the SEM will work with most customer environments.

As SEM deployments move beyond logging infrastructural events from routers, switches, servers, firewalls, and so forth, the ability to properly monitor business applications becomes crucial. Since most applications - especially those developed internally or by external software developers - do not include detailed logging it has become a challenge to incorporate this critical data into SEM products. Potential solutions to this challenge are based on network sniffing or other technologies.

SEMs are often sold to help satisfy U.S. regulatory requirements such as those of Sarbanes-Oxley, PCI-DSS, GLBA; in general the solutions these products can provide extend only to enhanced monitoring and analysis of enterprise computing activity; SEM is not a "magic bullet" for compliance but can be helpful in generating reports to support a limited set of controls.


One of the major problems in the SEM space is the difficulty in consistently analyzing event data. Every vendor, and indeed in many cases different products by one vendor, uses a different proprietary event data format and delivery method. Even in cases where a "standard" is used for some part of the chain, like Syslog, the standards don't typically contain enough guidance to assist developers in how to generate events, administrators in how to gather them correctly and reliably, and consumers to analyze them effectively.

As an attempt to combat this problem, a couple parallel standardization efforts are underway. First, The Open Group is updating their circa 1997 XDAS standard, which never made it past draft status. This new effort, dubbed XDAS v2, will attempt to formalize an event format including which data should be included in events and how it should be expressed. The XDAS v2 standard will not include event delivery standards but other standards in development by DMTF may provide a wrapper.

In addition, MITRE is also in the midst of a standardization effort called CEE that is somewhat broader in scope - it attempts to define an event structure as well as delivery methods.

Vendor products[edit]

See also[edit]


External links[edit]