Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

Version 1 Next »

Status: Active
Principal Author: Adam Slagell, CISO
Approved by: Dr. William T.C. Kramer, Blue Waters Deputy Director

Scope

This policy covers all users of NCSA HPC systems and Blue Waters Systems (test and development clusters, i.e. Blueprint, and Blue Waters HPC).

Introduction & Purpose

The primary threat to the security of NCSA production systems comes from user "identity theft", often exposed as compromised user accounts and credentials. Within the existing HPC environments managed by NCSA, 25% of security incidents stem from user credentials becoming compromised and used by unathorized persons for malicious purposes. This document details the policy covering the monitoring of user activity.

Policy Statement

The NCSA and Blue Waters support staff (e.g. system managers, Incident Response & Security Team (IRST)) monitors all NCSA and Blue Waters  production systems, in particular the HPC clusters and related resources. This extends to monitoring all user interactions with these resources including encrypted channels such as SSH. Secure communications channels may be modified to permit this monitoring to take place.

All users of Blue Waters and other HPC resources will be provided periodic notification that such monitoring is in place.

Process Details

Log Retention

The detailed SSH logs, which record most command line input and output as well as file transfers, are rotated out of use and discarded after approximately 4 weeks unless suspected activites have occurred. These logs provide much of the input for the host-based intrusion detection system. For specific security incidents, relevant portions of these logs may be saved into our incident response tracking system or other areas. Higher-level logs (e.g., network flows, IDS alerts, authentication logs, process accounting, and general system logs) may be held for longer periods of time.

Host-based Intrusion Detection System (HIDS)

Our intrusion detection/protection systems (IDPS) monitors the SSH commands executed and files down[up]loaded (as part of most everything processed though STDIN/STDOUT), looking for signs of account compromise. Users will be informed as soon as possible if inappropriate activites involving their account are detected.

  • No labels