Alert.png The wiki is deprecated and due to be decommissioned by the end of September 2022.
The content is being migrated to other supports, new updates will be ignored and lost.
If needed you can get in touch with EGI SDIS team using operations @

Difference between revisions of "EGI CSIRT:SMG"

From EGIWiki
Jump to navigation Jump to search
Line 33: Line 33:

*Pakiti client (eu.egi.sec.WN-Pakiti-ops)  
*Pakiti client (eu.egi.sec.WN-Pakiti-ops)  
**Reports a list of installed packages to the EGI Pakiti servers. The result of their evaluation is reported in the Pakiti columns of the dashboard.
**Reports a list of installed packages to the EGI Pakiti servers. The more detailed information about Pakiti results are available from the Dashboard.
**Error: Any critical vulnerability detected
**Error: Any critical vulnerability detected
** Warning: Any High-rated vulnerability detected
** Warning: Any High-rated vulnerability detected

Revision as of 11:19, 14 November 2011


public team pages| Incident Response Task Force (IRTF) | Security Drills Group (SDG) | Security Monitoring Group (SMG) |
public pages | Mission | Incident reporting | Dissemination | Alerts | Operational notices | Monitoring | Security challenges | Policies | Contacts |

Security Monitoring Group

Long-term Objectives

Security monitoring is a key component to security. The EGI CSIRT works to provide the EGI, NGI and site security staff with tools and procedures to detect and contain security incidents as well as weaknesses that could lead to a compromise. In doing so, the EGI CSIRT collaborates closely with the sites and NGI, however it does not provide replacements for site and NGI level monitoring. Even though the results developed are primarily ment for EGI, they are designed and developed so that they could be utilized on the site and/or NGI level as well.

The EGI CSIRT strives to collect various information from the infrastructure, even using own probes and sensors or by combining results generated by other systems (e.g., accounting) on different levels of the infrastructure. The data collected will be evaluated based on current needs and risks, and alarms raised accordingly. The security monitoring system will provide both high-level overview to get a quick notion about the infrastructure as well as a sufficient level of details necessary to sort out security issues detected. The system will archive and evaluate history to follow and forecast trends in security.

Security monitoring will also provide mechanisms and tools to assist inincident containment, for example to gather important log records. Information about users' activities will be used to identify last actions performed by a user to e.g. estimate the sites possibly infected. A way of (semi)automated processing of alerts and warnings issued by the EGI security groups will be investigate and possibly developed.

In order to accomplish the objectives, the EGI CSIRT will develop own tools, mainly to grid specific areas and use existing tools if possible and advise on their usage. Developed probes are not intrusive and do not attempt tocircumvent any security mehanisms and are not resource intensive. Results collected by these probes are only available to the EGI CSIRT members and communicated to the appropriate site security contacts. We will also provide recommendations and advisories to the sites and NGI as to how to combine existing technologies to provide efficient local security monitoring. When applicable, collaboration will be established with similar groups of NREN CSIRTs. The tools should be seamlessly integrated with common operational mechanisms used routinely by the security and operations.

Monitoring probes

The results produced by the security monitoring tools are available from the EGI Security dashboard, which records the outputs of the following probes:

  • CRL checks (eu.egi.sec.WN-CRL-ops)
    • Verifies that all CRLs exist and are current
    • ERROR: an out-dated or missing CRL detected
    • WARNING: an CRL is about to expire
  • Dangerous filesystem permissions (eu.egi.sec.WN-Permissions-ops)
    • Detects world-writable files or directories in exported environment. These files can be misused to information leak, tempering with results or even identity thefts.
    • ERROR: writable executables or directory with common commands
  • Vulnerable file permissions (eu.egi.sec.WN-FilePermVulns-ops)
    • Detects vulnerable file permissions as per CVE-2009-4033 (/var/log/acpid), which could lead to privilege escalations.
    • ERROR: Vulnerable file detected
  • Torque vulnerability (eu.egi.sec.WN-Torque-ops)
    • Checks whether torque server that the WN is using has vulnerable options turned on, as per EGI-ADV-20110615-02
    • ERROR: A vulnerable option detected
  • Pakiti client (eu.egi.sec.WN-Pakiti-ops)
    • Reports a list of installed packages to the EGI Pakiti servers. The more detailed information about Pakiti results are available from the Dashboard.
    • Error: Any critical vulnerability detected
    • Warning: Any High-rated vulnerability detected
  • Following probes are reported but will probably be removed (at least temporarily):
    • eu.egi.sec.ARCCE-Jobsubmit-ops
    • eu.egi.sec.CE-JobState-ops
    • eu.egi.sec.CREAMCE-JobSubmit-ops
    • eu.egi.sec.CE-JobSubmit-ops
    • eu.egi.sec.CREAMCE-JobState-ops

Any ERROR state are expected to get fixed within two businnes days. Any WARNINGs must be addressed within a month.

Vision for year 1

  • Help sites achieve a common baseline
    • Define (and document) mechanisms and best practices to gather logs and other data in orther to ease resolution of incidents and suspicious activities. Common repositories (e.g. central syslog) and tools for its management are inherently involved in this task.
  • Provide for (semi)automatic handling of alerts
    • CSIRT providing a machine-readable format of "interesting" characteristics (IP, user name, subject name, ...). Site-level tools being able to fetch these alerts and evaluated, e.g. using data from central syslog. On-demand and/or on-the-fly processing possible.
  • Further development of Pakiti and Nagios (some topics mentioned below)
  • Providing components as a service
    • We provide a basic suppport for independent installation of the tools, but will also explore the possibility to provide the functionality from our services (outsourcing)
    • (Simple) procedures describing manipulation with the sensitive data (passing among services, access rights)
  • A single point aggregating information available and providing a basic access to them - a dashboard


The lists of action at each task are not definite and will be discussed further. Some of tasks are well understood and there exist a vision about their focus. Other tasks will require more discussions about the actual contect and particular goals that are feasible.

Monitoring of security patches using Pakiti

  • finer access control, e.g., based on information from GOC DB
  • Generating statistics and reports
  • Tagging of and notifications (alarms) about severe vulnerabilities
  • Producing OVAL format for EGI advisories (e.g. SVG ones) and their processing
  • Support NGIs in setting a local instances of Pakiti
  • Performance improvements if needed/asked
  • Integration with existing monitoring frameworks (Nagios)
  • Pakiti server verification
  • Integration with dashboards, UI improvements
  • Operation of the EGI Pakiti server at CESNET
  • Simplification of the installation procedure

Security monitoring with Nagios

  • Development of new probes (based on risk analysis and experience with previous incidents)
  • Support for short-lived "dynamic" probes (if needed) (a procedure and template for a quick introduction of new "volatile" probes testing some very particular characteristics, which needs a fast reaction
  • guides, docs for NGIs/sites
  • Support NGIs in setting local instances of CSIRT Nagios
  • operation of the CSIRT Nagios instance at GRNET
  • Handling of results (raising alarms, sending notifications, access control to results, ...)
  • Evaluation of possible integration of security probes with standard Nagios (securing results, ...)
  • Integration with dashboards
  • Evaluation of results from multiple different sources
  • Aggregating security alerts in a single place

Tracing users

  • Tools to collect information about users from multiple sources (L&B, accounting, logs)
  • Evaluation of this data to trace users' last steps on demand
  • Evaluating log records produced by grid middleware (including checking that components logs appropriate information)

Site level tools

  • recommendation for setting up a central syslog server
  • specification of filters for the syslog
  • (semi)automatic processing of EGI security advisories (checking logs for IP addresses, DNs, ...)
  • best practises for log maintanence

Security Dashboard

  • providing a single place to overwiew current status, history and provide additional details
  • integration with existing EGI dashboards and appropriate frameworks will be evaluated



  • Daniel Kouril (, Czech Republic NGI


Name NGI Home Organization Effort Available (PM or FTE)
Stuart Kenny Ireland NGI TCD
David O'Callaghan Ireland NGI TCD
Christos Triantafyllidis Greek NGI AUTH
Jinny Chien - ASGC
Daniel Kouril Czech Republic NGI CESNET 0.25 FTE
Michal Prochazka Czech Republic NGI CESNET 0.25 FTE
Dusan Vudragovic Serbia NGI AEGIS
Angela Poschlad German NGI KIT
Bartlomiej Balcerek Poland NGI WCSS (CYFRONET) 4
Emir Imamagic Croatia NGI
Riccardo Brunetti Italy NGI INFN
Guiseppe Misurelli Italy NGI INFN
Dorine Fouossong France NGI