Passive monitoring of Modbus and DNP3 at solar, wind and BESS sites, without touching the device
How to watch Modbus and DNP3 traffic at a generation site without polling a single device, what to look for, and where CIP-015 does and does not apply.
Most solar, wind and battery sites run on two industrial protocols: Modbus TCP between the plant controller and the inverters, meters and battery management systems, and DNP3 between the site RTU or RTAC and the utility or the owner's operations center. Neither protocol was designed with authentication in mind in its common deployed form, and both will do what they are told by anything that can reach them on the network.
That makes the traffic itself the best record of what is happening on site. You can watch it without sending a single packet to a device. This post explains how, what is worth looking for, and where NERC's internal network security monitoring standard, CIP-015, does and does not apply to generation sites.
Why passive, and not polling
Active scanning tools send requests to devices to find out what they are. On an office network that is routine. On a plant network it is a risk: older inverters, meters and RTUs have small network stacks, and an unexpected request, or simply too many of them, can make a device drop its connection or reboot.
Passive monitoring only listens. It works from a copy of the traffic, so the devices never know it is there.
- A SPAN or mirror port on a managed switch copies the traffic from chosen ports to a monitoring port. Cheap, but a busy switch can drop mirrored packets, and the configuration lives on the switch.
- A network TAP is a small inline device that copies traffic at the cable. It does not drop packets under load, and it cannot send anything back onto the network.
- The sensor reads the copy, decodes Modbus and DNP3, and records who talked to whom, with which function codes, and when.
Where you place the copy matters more than the tool. The two places that see the most at a typical site are the switch behind the plant controller (inverter and BESS traffic) and the link between the RTU or RTAC and the outside world (DNP3 to the utility or operations center).
What normal looks like
The useful thing about a generation site is that its network is boring. The same plant controller polls the same inverters, at the same interval, with the same read requests, day after day. Once you have a few weeks of traffic you can write the baseline down:
- Which devices talk at all, and to whom (the plant controller to each inverter, the RTAC to the utility master).
- Which Modbus function codes appear. Reads (function code 3 or 4) will dominate; writes come from the plant controller when it sends setpoints.
- Which DNP3 function codes appear. Read requests and unsolicited responses are normal; controls should come only from the masters you expect.
- When things happen: polling cycles, the daily data pull, scheduled maintenance windows.
What is worth an alert
Against a quiet baseline, a short list of events covers most of what matters:
- A new talker. Any device that has never spoken Modbus (TCP port 502) or DNP3 (TCP port 20000 by default) on that segment before. A laptop plugged in by a technician shows up here.
- Writes from an unexpected source. Modbus write function codes (5, 6, 15, 16) from anything other than the plant controller. A write to an inverter's power limit register changes output.
- DNP3 controls from an unexpected master. Select and operate (3 and 4) or direct operate (5 and 6) arriving from an address that is not a known master.
- Restarts and stops. DNP3 cold or warm restart (13, 14) and stop application (18), and Modbus diagnostics (function code 8), which can put a device into listen-only mode. These are rare on a healthy site.
- Firmware and configuration sessions. Vendor tools often use their own protocols. A vendor-tool session outside a planned window is worth a look.
- Silence. A device that stops answering is either a field fault or something worse. Either way, someone should know.
Each alert needs an owner and a next step. A list of alarms that nobody reviews is not monitoring.
Where CIP-015 comes in
CIP-015-1 is NERC's internal network security monitoring (INSM) standard. Its stated purpose is "to improve the probability of detecting anomalous or unauthorized network activity in order to facilitate improved response and recovery from an attack."
Its scope is narrow. R1 covers networks protected by the Electronic Security Perimeter of high impact BES Cyber Systems and medium impact BES Cyber Systems with External Routable Connectivity. Exemption 4.2.3.6 says plainly that an entity with no BES Cyber Systems categorized as high impact or medium impact with External Routable Connectivity under CIP-002 is exempt. Most solar, wind and battery plants that fall under CIP at all are low impact, so CIP-015 does not apply to them. It does apply to an owner whose CIP-002 categorization produces medium impact systems with External Routable Connectivity, for example at a plant of 1,500 MW or more, or at an operations center that controls that much generation. Our CIP-002 categorization post covers how that is decided.
For entities in scope, the requirements line up with the approach above:
- R1.1: implement network data feeds, chosen with a risk-based rationale, to monitor "connections, devices, and network communications."
- R1.2: implement methods to detect anomalous activity using those feeds. The measures list "documentation of network communication baseline used to detect anomalous network activity" as one example of evidence.
- R1.3: implement methods to evaluate what was detected and decide further action; the measures mention escalation that could include the CIP-008 incident response plan.
- R2: retain the monitoring data tied to anomalous activity at least until the action is complete. Data unrelated to detected anomalies does not have to be kept.
- R3: protect the collected and retained monitoring data from unauthorized deletion or modification.
CIP-015-1 takes effect in the U.S. on October 1, 2028 for high impact BES Cyber Systems and medium impact systems at Control Centers and backup Control Centers, and on October 1, 2030 for other medium impact BES Cyber Systems with External Routable Connectivity. CIP-015-2 (Project 2025-02), written after FERC's Order No. 907 directed that monitoring be extended to Electronic Access Control or Monitoring Systems (EACMS) and Physical Access Control Systems (PACS), was approved by FERC on August 10, 2026. It adds those EACMS and PACS one year after each CIP-015-1 date: October 1, 2029 and October 1, 2031. Our CIP-015 guide keeps the details.
Why a low impact site might do it anyway
No standard requires a low impact solar farm to watch its Modbus traffic. There are still practical reasons to:
- Remote sites are unstaffed. Traffic is often the only witness to what happened on site between visits.
- It answers "what changed?" When an inverter's output drops, a record of writes to its registers separates a field fault from a configuration change.
- It checks the asset list. Every device that talks is on the network, whether or not it is on your inventory.
- It is a head start. If a future acquisition, expansion or recategorization brings medium impact systems, the baseline and the habit already exist.
Start small: one TAP behind the plant controller, a few weeks of capture, a written baseline, and a short alert list with a named reviewer.