Skip to main content
MainTegrity Cyber Security Framework (CSF)
MainTegrity Cyber Security Framework (CSF)
End-to-End Cyber Security for IBM z/OS

What is Early Warning?

MainTegrity CSF cyber security suite for z/OS CSF FoundationFIM+EarlyWarningNetWatchSupplyChain

Early Warning is the behavior-focused module of CSF. It catches attackers during reconnaissance and exploit attempts—before they achieve their objectives. It answers the question: "Is someone probing my system or trying to escalate privileges?"

Catch attackers in the act. Behavior detection at machine speed.

Why Early Warning Matters

Most attacks follow a predictable pattern:

Phase 1

Reconnaissance 

Attacker maps environment, tests permissions, looks for vulnerabilities

Phase 2

Exploit 

— Attacker leverages a gap to escalate privileges or gain unauthorized access

Phase 3

Attack

Attacker executes their objective (ransomware, data theft, sabotage)

Early Warning detects phases 1 and 2, enabling you to stop attackers before phase 3.

For certain exploit types — APF authorization escalation, RACF/ACF2/Top Secret privilege escalation, and privileged command execution — Early Warning detects the attempt before any damage is done.

    What Makes Early Warning Unique

    Early Warning Capabilities

    • Detect z/OS Authority / Security Tampering — Monitors for changes to security settings, exits, and audit configurations
    • Prevent UserID Impersonation — Detects and can block attempts to assume another user's identity
    • Stop Rogue Encryption in Seconds — Identifies ransomware encryption patterns and suspends immediately
    • Alert on Suspicious User Behavior — Tracks unusual access patterns, privilege escalation, and reconnaissance

    Dynamic Authority Change Detection

    ⚠️ Claim requires engineering validation before external use.
    Early Warning is believed to be the only product in the world that can dynamically identify an authority change in a running z/OS environment. 

    In a normal production system, authority changes should never occur at runtime. Early Warning detects them as they happen and can block them immediately.

    60-Second Response Window

    Early Warning checks for threshold violations every 60 seconds by default (configurable). While there is a brief lag between when a malicious program starts and when the offending task is stopped, this is dramatically faster than conventional approaches.


    Result: Damage is limited to a few seconds, rather than the hours or days it may take to recognize and halt an attack with existing tools.

    Covers 42 Unique Monitoring Items

    Early Warning monitors up to 42 unique items for RACF implementations, with similar coverage for ACF2 and Top Secret.

    Access pattern monitoring includes:

    • Protected datasets and operator commands
    • BPX resources and FACILITY class
    • JES Spool resources and SDSF functions
    • ICSF functions and certificate resources

    Continuous Dataset Monitoring

      Early Warning monitors active data files and libraries for unusual actions — even if they are not included in regular validation scans.  Accesses are collected for an interval, and if they exceed customer-defined historical trends, an alert or other action can be triggered.

    Extended Tracking

      Early Warning monitors activities beyond normal z/OS capabilities. Attackers plan reconnaissance to avoid activities the system normally records — Early Warning catches them doing things they wouldn't expect to be tracked.

    Countermeasures: Stop Attacks, Don't Just Alert

    Early Warning can automatically block certain attack techniques:

    Attack TypeAvailable ActionsCountermeasure Result
    Program authorization escalationSuspend, Cancel, CounterSystem 47 abend — privileged operation blocked
    Privilege escalationSuspend, Cancel, CounterCountered — user remains unprivileged
    User impersonationSuspend, Cancel, CounterCountered — impersonation attempt fails
    Privileged command usageSuspend, CancelTask suspended or cancelled
    APF library updateCounterUpdate blocked
    Configuration / STC library updateCounterUpdate blocked
    Attack patterns (Layer 3)Suspend, CancelOffending task stopped

    When countermeasures are enabled, attackers receive cryptic system errors while Early Warning alerts your team.

    Three Layers of Detection

    Layer 1

    Reconnaissance Detection

    Monitors aggregate counts of suspicious activities over a configurable polling interval. Alerts when customer-defined thresholds are exceeded.

    Excessive Read Accesses To:

    • Authorized (APF) libraries
    • PARMLIBs, PROCLIBs
    • VTAMLST, TCPPARMS
    • Datasets used in started tasks

    User Access Checking Of:

    • Protected datasets and operator commands
    • BPX resources and FACILITY class resources
    • XFACILIT class resources
    • JES Spool and SDSF functions
    • ICSF functions and certificate resources
    • Program resources and TapeVol resources

    Special RACF Commands:

    • SETROPTS LIST
    • LU * (List User wildcard)
    • SEARCH, LISTDSD, RLIST

    Special Top Secret Commands:

    • TSS MODIFY, TSS LIST(ACIDS), TSS WHOHAS

    Excessive Display Commands:

    • MVS, JES, VTAM, and other display commands

    When thresholds are exceeded:

    1. Immediate alert that thresholds were passed
    2. Summary report detailing all reconnaissance activity

    Layer 2

    Exploit Detection

    Certain exploit attempts are intercepted before the action completes. Immediate actions vary by item type.

    Suspend + Cancel + Counter:

    • Program authorization escalation (non-authorized program becomes authorized)
    • User privilege escalation (user gains RACF SPECIAL or equivalent)
    • User impersonation (assuming another user's identity)

    Suspend + Cancel:

    • Privileged operator command usage (SETPROG, SET, VARY, and client-definable commands)

    Counter only:

    • Update to authorized (APF) library
    • Update to configuration / STC library

    Detection and reporting only:

    • Change to security audit options
    • Changes to TSO authorized command tables
    • Changes to TSO authorized program tables
    • Dynamic changes to APF list
    • Dynamic changes to system exits
    • Attempted logons to IBMUSER or ACFUSER
    • Excessive failed logons in a period

    Layer 3

    Attack Detection

    Monitors for attack signatures during the polling interval. Triggers alerts and optional immediate actions (Suspend / Cancel) when customer-defined thresholds are exceeded.

    • Scanning excessive files/datasets
    • Updating excessive files/datasets
    • Deleting excessive files/datasets
    • Hardware encryption while updating excessive files (ransomware signature)
    • Excessive hardware encryption instructions
    • High CPU usage combined with high I/O

    Technical Q&A: Use Cases

    Use Case 1: Insider Privilege Escalation

    Scenario: A system programmer attempts to grant themselves RACF SPECIAL authority.

    Early Warning Detection:

    • Layer 2 (Exploit Detection) triggers on the privilege escalation attempt
    • Countermeasure blocks the escalation — user remains unprivileged
    • Attacker receives a cryptic System 47 abend

    Response:

    • Alert sent to security team with user details
    • Forensics browser shows exact command attempted
    • User ID revoked pending investigation

    Result: Escalation blocked entirely; no privileges gained.

    Use Case 2: Reconnaissance Detection

    Scenario: A compromised account begins mapping the z/OS environment.

    Early Warning Detection:

    • Layer 1 (Reconnaissance) detects excessive RACF SEARCH commands
    • Read accesses to APF libraries exceed threshold
    • SETROPTS LIST commands indicate environmental mapping

    Response:

    • Immediate alert when thresholds exceeded
    • Summary report details all reconnaissance activity
    • Account suspended before exploit phase begins

    Result: Attacker detected during preparation; attack prevented.

    Use Case 3: Ransomware Encryption Pattern

    Scenario: Ransomware begins encrypting datasets using hardware encryption.

    Early Warning Detection:

    • Layer 3 (Attack Detection) identifies encryption combined with excessive file updates
    • Hardware encryption activity flagged as ransomware signature
    • Threshold exceeded within 60 seconds

    Response:

    • Offending task suspended automatically
    • Alert sent with affected datasets list
    • FIM+ Restore Assist guides recovery

    Result: Encryption stopped after ~60 seconds of damage; rapid recovery.