← All posts
CIP-009-6 · CIP-011-3 · CIP-002-5.1a

CIP-009 recovery plans: backing up PLC, RTAC and inverter configurations

What a CIP-009-6 recovery plan must cover, part by part, and how to back up, verify and store PLC, RTAC, relay, inverter and firewall configurations.

When a controller fails or a cyber incident forces a rebuild, the question is not whether you have a backup. It is whether the backup is current, whether it restores onto the hardware and firmware you run today, and whether the person on shift can find it. CIP-009 turns those questions into written requirements for BES Cyber Systems.

This post covers who CIP-009-6 applies to, what each part asks, and the practical work behind it at a solar, wind or battery site: what to back up for PLCs, RTACs, relays, inverters, BESS controllers and firewalls, how to prove a backup restores, and what evidence to keep.

Does CIP-009 apply to you?

CIP-009-6 applies to high and medium impact BES Cyber Systems and their associated Electronic Access Control or Monitoring Systems (EACMS) and Physical Access Control Systems (PACS). Section 4.2.3.5 of the standard exempts Responsible Entities that have no BES Cyber Systems categorized as high or medium impact.

  • Most solar, wind and battery plants that are under CIP rate low impact. Under CIP-002-5.1a criterion 2.1, a plant is medium impact only when its generation at a single plant location reaches 1,500 MW or more in a single Interconnection.
  • A Control Center or backup Control Center that performs Generator Operator functions for 1,500 MW or more in a single Interconnection is medium impact under criterion 2.11. A fleet run from one control room can cross that line even when every plant is small.
  • Category 2 inverter-based resources are not subject to the CIP standards. Registering one with NERC does not bring CIP-009 with it.

If all your BES Cyber Systems are low impact, CIP-009 is not an obligation, and you will not be audited against it. A recovery plan is still good practice: the work below is what gets a site generating again after a failed controller, whether or not a standard asks for it.

Medium impact: the plant and the Control Center are scoped differently

Read the Applicable Systems column for each part. Several parts apply only to high impact systems and to medium impact systems at Control Centers, so a medium impact system at a generating plant owes less than one in a control room.

R1: what the recovery plan must contain

R1 requires one or more documented recovery plans that collectively include:

  • Conditions for activating the plan (Part 1.1).
  • Roles and responsibilities of responders (Part 1.2).
  • One or more processes for the backup and storage of information required to recover BES Cyber System functionality (Part 1.3).
  • One or more processes to verify that the backups in Part 1.3 completed successfully and to address any backup failures (Part 1.4). This part applies to high impact systems and to medium impact systems at Control Centers, with their EACMS and PACS.
  • One or more processes to preserve data, per Cyber Asset capability, for determining the cause of a Cyber Security Incident that triggers the plan. The standard adds: "Data preservation should not impede or restrict recovery" (Part 1.5).

Parts 1.1, 1.2, 1.3 and 1.5 apply to every high and medium impact BES Cyber System and its associated EACMS and PACS, including medium impact systems at a plant.

R2: testing the plan and the backups

R2 applies to high impact systems and to medium impact systems at Control Centers, with their EACMS and PACS. It does not apply to medium impact systems that are not at a Control Center.

  • Test each recovery plan at least once every 15 calendar months: by recovering from an actual incident, with a paper drill or tabletop exercise, or with an operational exercise (Part 2.1).
  • Test a representative sample of the information used to recover BES Cyber System functionality at least once every 15 calendar months, to ensure it is usable and compatible with current configurations. An actual recovery that uses that information substitutes for the test (Part 2.2).
  • High impact only: test each plan at least once every 36 calendar months through an operational exercise in an environment representative of production. An actual recovery response may substitute (Part 2.3).

R3: updating the plan

R3 has the same scope as R2: high impact systems and medium impact systems at Control Centers.

  • No later than 90 calendar days after a recovery plan test or an actual recovery: document the lessons learned, or that there were none; update the plan based on them; and notify each person or group with a defined role in the plan of the updates (Part 3.1).
  • No later than 60 calendar days after a change to roles or responsibilities, responders, or technology that you determine would affect your ability to execute the plan: update the plan and notify each person or group with a defined role (Part 3.2).

Replacing a controller model, moving to a new engineering software version or changing the backup tool can each be a technology change. Decide in advance which changes you treat as affecting the plan, and write that rule down.

What to back up, device by device

Part 1.3 asks for the information required to recover functionality. For OT devices that is usually more than a configuration file. For each device, the backup should let someone rebuild it on a replacement unit without guessing.

  • PLCs: the project file from the engineering software, not just an upload from the controller (an upload can lose comments and tag descriptions), plus the firmware version, the module list and any network or communication settings held outside the project. Rockwell Studio 5000 (.ACD files), Siemens TIA Portal and Schneider EcoStruxure Control Expert all keep the project on the engineering workstation.
  • RTACs: the SEL AcSELerator RTAC project, the firmware version it was built for, and the device's user accounts, certificates and network settings.
  • Protective relays: the settings files from the vendor tool (AcSELerator QuickSet for SEL relays), with the relay part number and firmware version, since settings are tied to both.
  • Inverters: the parameter set and grid-code settings from the vendor's commissioning software or web interface, plus the firmware version and communication settings (Modbus address map, IP addressing). Note who holds them: on many sites the inverter vendor or EPC commissioned the units and holds the only copy.
  • BESS controllers and power plant controllers: the controller configuration and logic, setpoints, and the integration settings between the BMS, PCS and plant controller, which often come from different vendors.
  • Firewalls and switches: the full running configuration, exported in a format the device can import, with the firmware version.
  • Servers, HMIs and engineering workstations: a system image or a documented rebuild procedure, the installers and license files for the engineering software at the right versions, and the HMI project.
  • The documents people need under pressure: network diagrams, IP plans, and where the credentials needed for recovery are kept.

Take a backup after every approved change, not only on a schedule. A backup from before the last change restores the wrong configuration, and Part 2.2 tests for exactly that: compatibility with current configurations.

Know where vendor tools keep configurations

Most OT configurations live in two places at once: in the device, and in a project file on an engineering workstation or a vendor laptop. Recovery usually depends on the project file. List every device, the tool that configures it, the tool version, and where the authoritative project file is kept. If the only copy is on a contractor's laptop, your Part 1.3 process does not cover it yet; ask for the files to be handed over after each site visit, and write that into the contract.

Prove the backup restores

A file that exists is not a backup that works. The Part 2.2 measures give examples of evidence: test results with criteria for usability and for compatibility with current system configurations, such as a manual or automated comparison between the backup contents and the current configuration. Practical ways to do that:

  • Open each sampled project in the engineering software version you run today, and confirm it loads without conversion errors.
  • Compare the backup with what is on the device now: the firmware version, and for network devices a line-by-line comparison of the configuration.
  • Restore onto a spare or a test bench unit where you have one, and record the result.
  • Rotate the sample so it covers each device type and vendor tool over time, not the same easy device every year.

Write down the criteria before the test, record the result, and fix failures with a dated note. For high impact systems and medium impact systems at Control Centers, Part 1.4 also requires a process to confirm that backups completed and to address failures.

Store backups as BES Cyber System Information

A device configuration usually shows IP addresses, firewall rules, accounts and network layout: the kind of information an entity's CIP-011 program typically identifies as BES Cyber System Information (BCSI). For high and medium impact systems, CIP-011-3 R1 requires methods to identify BCSI (Part 1.1) and methods to protect and securely handle it to mitigate risks of compromising confidentiality (Part 1.2). Treat the backup repository as a BCSI storage location: access limited to the people who need it, and the location named in your information protection program.

Keep at least one copy where an incident on the plant network cannot reach it, such as offline media or a separate security zone, and record a hash of each backup so you can show later that it has not changed.

Evidence to keep

  • The dated recovery plan, showing the activation conditions, the roles, and the backup, verification and data preservation processes.
  • The device list with each device's backup location, tool and version, and the date of its last backup.
  • Backup completion records and how failures were addressed (Part 1.4, where it applies).
  • Dated test records for each plan, and the backup sample tests with their criteria and results (R2, where it applies).
  • Lessons learned or a dated note that there were none, the revised plan, and proof that the people with roles were notified (R3, where it applies).

CIP-009-6 asks you to retain evidence of each requirement for three calendar years, unless your Compliance Enforcement Authority directs you to keep it longer as part of an investigation.

Where automated evidence collection helps

A collector cannot back up PLC logic or rebuild a device. What it can give you is an independent, dated record of each device's configuration as it was collected. For the devices it reads, that record is a reference to compare a backup against before you trust it, and comparing each collection with the device's approved baseline shows when a change has made the last backup out of date.