EtherCAT's base wire protocol — standardized by the EtherCAT Technology Group as ETG.1000 and internationally as IEC 61158 Type 12 — has no authentication, encryption or message-integrity field of its own; every device wired into the segment can, by design, read and modify the single Ethernet frame as it streams past in hardware. The EtherCAT Technology Group does not contest this at the datagram level; its current security position instead argues the protocol's closed, IP-free, hardware-processed architecture limits what an attacker can do, and that EtherCAT systems have been independently assessed by UL Solutions as meeting IEC 62443-3-3 Security Level 2 without protocol changes. Academic research from 2017-2020 demonstrated that a device with physical access to the bus can still inject and manipulate frames, and the one CISA advisory that actually touches EtherCAT covers a third-party traffic-analysis tool, not the protocol or a master/slave implementation.
Key takeaways
- EtherCAT is a real-time industrial Ethernet fieldbus in which one MainDevice (master) sends a frame that passes through every SubDevice (slave) in turn, and each SubDevice reads and writes its data while the frame moves.
- No, not in the base datagram format; the security case ETG makes is architectural.
- UL Solutions certified a reference EtherCAT system, not individual devices, against three threat scenarios.
- Mailbox protocols carry acyclic data next to the cyclic process data, and two of them reach beyond the fieldbus.
What is EtherCAT, and how does "processing on the fly" work?
EtherCAT is a real-time industrial Ethernet fieldbus in which one MainDevice (master) sends a frame that passes through every SubDevice (slave) in turn, and each SubDevice reads and writes its data while the frame moves. The ETG technology page puts it this way: "Each EtherCAT SubDevice reads the data addressed to it 'on the fly', and inserts its data in the frame as the frame is moving downstream." The last device reflects the frame back to the MainDevice. Frames carry EtherType 0x88A4 directly in Ethernet, with no IP header.
The read/write happens in a dedicated chip on each SubDevice, the EtherCAT Slave Controller (ESC), not in the device CPU or its software. That is what buys EtherCAT its motion-control determinism; ETG quotes cycle times of 100 µs or less and jitter of 1 µs or less. It is also why EtherCAT compares poorly with Modbus or PROFINET as a general network: it is built to move process data, not to route traffic.
The specification has two names. ETG maintains the ETG.1000 series (current release 1.0.4, 15 September 2017), and IEC publishes the same technology as Type 12 in IEC 61158, for example IEC 61158-6-12:2019, edition 4.0.
| ETG.1000 part | Content | IEC counterpart |
|---|---|---|
| ETG.1000.1 | Overview | — |
| ETG.1000.2 | Physical layer service and protocol | IEC 61158-2 (Type 12) |
| ETG.1000.3 / .4 | Data link layer service / protocol | IEC 61158-3-12 / -4-12 |
| ETG.1000.5 / .6 | Application layer service / protocol | IEC 61158-5-12 / -6-12 |
None of these parts defines a credential, key or signature field.
Does EtherCAT have built-in authentication or encryption?
No, not in the base datagram format; the security case ETG makes is architectural. Its cybersecurity page lists a "clear communication hierarchy: Frames can only be generated by the MainDevice", "No IP-based communication – protection against typical malware attacks", and frame handling in hardware. ETG's April 2026 whitepaper adds that "The ESC is not capable of sending frames itself."
Those points narrow the attack surface; they do not authenticate anything. ETG's own assessment notes, covered in the next section, mark the IEC 62443-3-3 requirements for public-key authentication (SR 1.9) and use of cryptography (SR 4.3) as not applicable in its first two reference scenarios. Cryptographic pairing appears only in the third scenario, as a software extension. For applications with very high requirements, the cybersecurity page says EtherCAT "is extended with additional security features such as authenticated, quantum-secure encryption in a backward-compatible manner". Read that as an extension a vendor has to implement, not a property of ETG.1000. The NIST SP 800-82 OT security guide treats this architecture-over-cryptography pattern as typical of real-time fieldbuses.
What did UL Solutions certify against IEC 62443-3-3?
UL Solutions certified a reference EtherCAT system, not individual devices, against three threat scenarios. ETG's Explanatory Notes of 20 April 2026 state that "all 103 system requirements (SR) from the IEC 62443-3-3 standard were evaluated", list certificates DK-177530-UL, DK-178394-UL and DK-178399-UL, and say certification "does not require certification of the EtherCAT devices used."
| Blueprint scenario | Attacker / need | What is added to plain EtherCAT |
|---|---|---|
| 1: machine unit in a typical factory | SL 2 | Nothing; built-in features and configuration rules |
| 2: less restrictive access control | up to SL 3, no confidentiality need | Software features in the MainDevice only (propagation-delay intrusion check, parameter checks, ETG-signed ESI files, device inventory check, signed firmware files) |
| 3: confidentiality required | up to SL 3 | Symmetric-key pairing of MainDevice and SubDevice, cryptographic integrity and confidentiality, "software extensions only; no changes to the EtherCAT chips (ESC)" |
The specification for those extensions is ETG.1800 EtherCAT Security Profile. Version 0.9.0 was published to ETG members as a draft for review on 29 September 2026, with a review period that ended on 11 November 2026. Part 1 covers the system and blueprint scenarios; Part 2 covers the EtherCAT measures for IEC 62443-3-3 assessments. Our IEC 62443-3-3 guide explains what system SL claims mean, and the IEC 62443-4-2 guide covers the separate component certificate a device maker may still need.
Which mailbox protocols open a path into the segment?
Mailbox protocols carry acyclic data next to the cyclic process data, and two of them reach beyond the fieldbus. The technology page says "TCP/IP connections can optionally be tunneled through a mailbox channel", and describes File access over EtherCAT as "similar to TFTP" for "a uniform firmware upload to devices across a network".
| Protocol | What it carries | Security relevance | ETG control in the Explanatory Notes |
|---|---|---|---|
| CoE (CANopen over EtherCAT) | Object dictionary access | Parameters can be changed | MDR08/MDR25: rewrite and check parameters |
| EoE (Ethernet over EtherCAT) | Tunnelled Ethernet/IP traffic | Brings IP services (web server, FTP) to field devices | MDR03: EoE disabled by default in the MainDevice; SDR01 |
| FoE (File access over EtherCAT) | Files, mainly firmware | Firmware replacement path | SDR03; integrity check before applying; signed firmware file in Scenario 2 |
| SoE / AoE | Drive profile / ADS services | Device commands | Restrict unneeded functions (MDR15) |
| Mailbox Gateway, UDP EtherCAT | Mailbox or frames over UDP/TCP port 0x88A4 | Reaches the segment from an IP network | MDR21–MDR23: disabled by default, block port 0x88A4 |
The whitepaper is direct about EoE: tunnelled traffic "can compromise the security of inadequately protected field devices at the end of the tunnel", and EoE "can be disabled either globally or selectively within the MainDevice."
Is Safety over EtherCAT a security feature?
No. Safety over EtherCAT (FSoE) protects against random transmission errors for functional safety; it is not designed against a deliberate attacker. ETG's safety page says FSoE was developed according to IEC 61508, is standardized in IEC 61784-3, and is suitable up to SIL 3, and that "the communication system is part of a Black Channel, which is not considered to be safety-relevant." The IEC text is IEC 61784-3-12, consolidated edition 1.1 (2019).
The practical consequence: a safety assessment of an FSoE device does not answer a security question. A SubDevice maker still needs separate security requirements, and IEC 62443-4-2 is where those sit.
What does independent research show about attacking a real EtherCAT segment?
A device with access to the physical bus can inject and manipulate frames; no study has shown a remote attack on a correctly isolated segment. Granat, Höfken and Schuba of FH Aachen (CNSCE 2017) found EtherCAT "lacks common security parameters such as authentication, encryption, and authorization" and built ecat_mitm to rewrite datagram payloads in transit. They also noted that IP-layer tricks such as ARP spoofing do not apply: an attacker needs a cable tap, a free port, or a compromised device in the ring.
Kevser Ovaz Akpinar and Ibrahim Ozcelik (Security and Communication Networks, 2018) reached the same conclusion, found the protocol open to MAC spoofing, MITM, denial-of-service and command injection, and proposed a Snort-based preprocessor for detection. ETG's MainDevice requirements read as answers to exactly these findings: close unused ports (MDR06), "Auto Close" used ports (MDR07), count frames with the datagram index field (MDR17) and time the frame's return path (MDR16, MDR24).
Which EtherCAT vulnerabilities have CVE IDs?
The published CVEs are in software that processes EtherCAT traffic: master stacks, drivers and parsers, not the ESC chips. As of 14 November 2026:
| CVE | Component | Score (CVSS 3.1) | Fix |
|---|---|---|---|
| CVE-2026-70388 | SOEM ≤ 2.0.0: integer underflow in mailbox length check, out-of-bounds write | 8.3 (CNA) | SOEM 2.0.1 |
| CVE-2026-82944 | SOEM: multi-segment transfer overflow | 8.8 (CNA) | SOEM 2.0.1 |
| CVE-2026-82943, -82945, -82946, -82947 | SOEM: out-of-bounds reads, FMMU array overflow, null pointer | 6.5–7.1 (CNA) | SOEM 2.0.1 |
| CVE-2023-7243, CVE-2023-7244 | CISA ICSNPP EtherCAT Zeek plugin ≤ d78dda6: out-of-bounds write | 9.8 (CISA, NVD) | Commit 3bca34c or later |
| CVE-2023-7242 | Same plugin: out-of-bounds read | 8.2 (CISA, NVD) | Commit 3bca34c or later |
| CVE-2020-12494 | Beckhoff TwinCAT RT driver for Intel 8254x/8255x: memory leaked in frame padding ("EtherLeak") | 5.3 (CERT@VDE, which NVD labels Secondary); NVD's Primary score is CVSS v2 5.0 | Per VDE-2020-019 |
The SOEM set matters most to device makers, because SOEM is a common open-source MainDevice stack in embedded controllers. The rt-labs advisory RRTL-260921-01 says all six "can be triggered by a malicious or compromised device on the same EtherCAT network", through the SoE, CoE, EoE or FoE handlers. SOEM 2.0.1 shipped on 15 October 2026. So the trust model runs both ways: a rogue SubDevice can attack the master.
The only CISA ICS advisory naming EtherCAT is ICSA-24-051-02 (20 February 2024), for the ICSNPP EtherCAT plugin. Beckhoff's CISA advisories, most recently ICSA-24-312-01 (CVE-2024-8934, TwinCAT Package Manager, 7 November 2024), concern engineering software, not EtherCAT.
What should a SubDevice or MainDevice maker build in?
Build the controls ETG's assessment relies on into the product, so integrators can switch them on. A practical list, mapped to the Explanatory Notes:
Support explicit device identification (SDR02), so a MainDevice can detect a swapped or foreign device. Refuse firmware upload unless intended (SDR03), and check firmware integrity before applying itFor Scenario 2, ship firmware files with a certificate verifiable from the ESI file; our OTA update security guide covers signing. Harden any EoE application (SDR01)A web server reached through EoE needs the same authentication and patching as one on a LAN. Protect authenticators in hardware (SDR04)if you target Scenario 3. For MainDevice stacks: validate every mailbox length fieldThe SOEM CVEs came from trusting slave-supplied lengths. Ship EoE and Mailbox Gateway disabled by default (MDR03, MDR21), drop frames the master did not send (MDR09), and log segment events with timestamps (MDR12). Track the stack as a componentRecord SOEM, a vendor master stack or ESC firmware with its version in your firmware SBOM so the next advisory maps to products in minutes.
Worked example: how would you harden a synthetic I/O module before release?
Synthetic example; the device and numbers are invented for illustration. A maker of a 16-channel EtherCAT digital I/O module with an embedded web page for diagnostics reviews release 3.2 against the Explanatory Notes:
| Finding | Requirement | Action for release 3.3 |
|---|---|---|
FoE accepts any file named *.efw |
SDR03, integrity check | Verify a signature before flashing; reject unsigned files |
| Diagnostic web page over EoE has a fixed password | SDR01 | Per-device password set at commissioning |
| No explicit device ID | SDR02 | Implement explicit device identification |
| Gateway variant bundles SOEM 2.0.0 | CVE-2026-70388 and five others | Move to SOEM 2.0.1; list it in the SBOM |
The module needs no protocol change; every action is firmware and documentation.
What should an integrator check first?
Treat physical access as the credentialLock cabinets, document ports, and treat an unknown device on the ring as an incident. Keep EtherCAT off routed networksBlock UDP and TCP port 0x88A4 at the controller and leave Mailbox Gateway and EoE off unless needed. Map your system to a Blueprint Scenariofrom your own risk assessment; the UL certificates cover ETG's scenarios, not your machine. Patch tooling around the segment, including master stacks and Zeek-based monitoring. Inventory every MainDevice and SubDevice with firmware versions
What are the common EtherCAT security mistakes?
- Reading "SL 2 certified" as a device certificate. It is a system certificate for ETG's reference scenarios.
- Treating FSoE as a security control.
- Leaving EoE enabled after commissioning because a diagnostic web page once needed it.
- Accepting any file over FoE.
- Assuming SubDevices are trusted inputs to the master stack. SOEM shows otherwise.
- Waiting for an EtherCAT-specific CISA advisory before acting; the research above already shows what bus access allows.
Where does SoftWSP fit?
SoftWSP works on the IP side and on firmware, not on the EtherCAT wire. OT Snapshot reads Zeek logs from a mirror port on the routed network. It lists hosts, ports and remote-access paths, and it can show traffic to port 0x88A4 that crosses that network. It does not decode EtherCAT frames inside the cell. For device makers, our firmware security service builds a CycloneDX SBOM from a firmware image and matches components such as SOEM against NVD, CISA KEV, FIRST EPSS and CISA ICS advisories. AI drafts a cited report and an engineer reviews every finding. See OT Snapshot for the network side.
This guide draws on ETG's technology, cybersecurity, safety and specification pages, its April 2026 whitepaper and Explanatory Notes, IEC webstore records, the rt-labs SOEM advisory, CERT@VDE, CISA's CSAF records for ICSA-24-051-02 and ICSA-24-312-01, NIST SP 800-82, and the two academic studies, all read on 2026-11-14.
Put this guide to work
Frequently asked questions
Does EtherCAT have built-in authentication or encryption in its wire protocol?
No. EtherCAT's base protocol, standardized by the EtherCAT Technology Group (ETG) as the ETG.1000 series and internationally as IEC 61158 Type 12 / IEC 61784-2 Communication Profile Family 12, has no authentication field, no encryption and no message-integrity check anywhere in its application-layer datagram format. The protocol's defining mechanism, 'processing on the fly,' is described on ETG's own technology page as each SubDevice (slave) reading the data addressed to it 'on the fly, and insert[ing] its data in the frame as the frame is moving downstream' — the MainDevice (master) emits a single Ethernet frame that streams through every device on the segment in hardware, with the EtherCAT Slave Controller (ESC) doing the read/write entirely at the hardware level, independent of any device's CPU, operating system or software stack. Nothing in that datagram format carries a credential, a session key, or a signature; any device wired into the segment, by design, can read and write the frame as it passes. ETG does not dispute this at the protocol level — its own current cybersecurity messaging instead argues the architecture itself narrows what an attacker can do, not that the wire format carries cryptography.
If EtherCAT has no native authentication, how does the EtherCAT Technology Group justify calling it secure?
ETG's current cybersecurity page (fetched directly for this guide) argues EtherCAT is 'inherently robust against cyberattacks' for architectural reasons rather than cryptographic ones: communication is tightly controlled because 'frames can only be generated by the MainDevice,' the protocol carries no IP stack (so typical IP-based malware and routing attacks do not apply directly to the fieldbus segment itself), and frame processing happens in dedicated hardware rather than a general-purpose software stack, which the page says makes many attack classes, including denial-of-service, largely ineffective against a correctly configured segment. Building on that architectural baseline, ETG states EtherCAT systems have been 'assessed and certified by independent testing bodies (including UL Solutions)' as meeting 'Security Level 2 according to IEC 62443-3-3 — without any modifications to the protocol,' and that for 'higher security requirements (Security Level 3), only software extensions are necessary.' ETG also publishes a dedicated specification, ETG.1800 ('EtherCAT Security Profile'), whose draft parts 1 and 2 were released for review, aimed at guiding integrators and device makers on applying those software-layer security measures. None of this changes the underlying fact confirmed above: the base ETG.1000 datagram format itself still carries no authentication or encryption — the SL2 claim is about a system-level architecture and compensating controls evaluated against IEC 62443-3-3's roughly 103 system requirements, not a claim that the wire protocol was retrofitted with cryptography.
What does independent security research show about attacking a real EtherCAT segment?
Multiple academic studies, working independently of each other between 2017 and 2020, demonstrated that a device with access to the physical EtherCAT bus can inject and manipulate traffic, consistent with ETG's own acknowledgment that the protocol relies on controlling who reaches the wire rather than on authenticating what reaches it. Granat, Höfken and Schuba (FH Aachen University of Applied Sciences), presented at the 2nd International Conference on Computer, Network Security and Communication Engineering (CNSCE 2017), found EtherCAT 'lacks common security parameters such as authentication, encryption, and authorization' and built a tool called ecat_mitm to demonstrate frame interception and data manipulation — for example, rewriting a specific datagram's payload value in transit. Their paper makes an important distinction: because EtherCAT datagrams are encapsulated directly in raw Ethernet frames rather than routed IP traffic, classic IP-layer MITM techniques like ARP spoofing do not work against it — an attacker instead needs physical access to the bus (a cable tap) or control of a device already participating in the EtherCAT ring. Kevser Ovaz Akpinar and Ibrahim Ozcelik's 2018 paper in Security and Communication Networks (Wiley), describing itself as the first research specifically targeting EtherCAT vulnerabilities, independently reached the same conclusion — that EtherCAT 'does not have any flow or connection-based security to recognize master and slaves,' making it susceptible to MAC spoofing, MITM, denial-of-service and command-injection — and proposed a Snort-based trust-node preprocessor as a compensating detection layer. A separate 2020 formal-methods paper (Feng Tao et al., Journal of Computer Research and Development) identified three categories of MITM vulnerability in the protocol's message exchange: tampering, replay and deception. None of these findings breaks a cryptographic control, because the base protocol does not implement one at the datagram level — they demonstrate what unauthenticated frame processing allows once an attacker is on the segment.
Has CISA published an ICS advisory specifically about EtherCAT, and what should a device maker or integrator check first?
No CISA ICS advisory describes a vulnerability in the EtherCAT protocol itself or in a Beckhoff-built EtherCAT master/TwinCAT stack parsing EtherCAT traffic — the closest confirmed match is CISA's ICSA-24-051-02 (released 20 February 2024), which covers the open-source ICSNPP EtherCAT plugin for Zeek, a third-party network-monitoring tool used to analyze EtherCAT traffic for intrusion detection, not an EtherCAT master, slave or device firmware. That advisory assigns three CVEs against plugin versions d78dda6 and prior — CVE-2023-7244 and CVE-2023-7243 (both CWE-787, Out-of-Bounds Write, CVSS v3.1 9.8) and CVE-2023-7242 (CWE-125, Out-of-Bounds Read, CVSS v3.1 8.2) — all triggered by a specially crafted EtherCAT packet reaching the Zeek parser; CISA's advisory states the fix is updating to commit 3bca34c or later. Separately, CISA's archive contains several Beckhoff advisories (most recently ICSA-24-312-01, covering CVE-2024-8934, an OS command injection in TwinCAT Package Manager that a local user with administrative access rights can trigger by entering crafted values in its settings, CVSS v3.1 6.5 per CISA and CERT@VDE VDE-2024-064), but none of those found relate to the EtherCAT wire protocol or its security model either — they are unrelated software bugs in Beckhoff's engineering tooling. Given that gap, an integrator's practical checklist should rest on the architecture and the academic research above rather than wait for a protocol-specific advisory: inventory every EtherCAT MainDevice and SubDevice on the segment and confirm who has physical access to the cabling and any unused ports, since the academic attacks above all assume bus access as the starting point; verify whether the installed EtherCAT stack and any connected engineering software (TwinCAT or otherwise) are current against CISA's general advisory archive for that vendor; treat the EtherCAT segment itself as a trusted-physical-access zone, not a network a remote attacker should ever reach, and keep it off any routed IP network since the protocol was never designed to be exposed that way; and if IEC 62443-3-3 Security Level 2 or 3 compliance matters for the deployment, confirm which specific ETG.1800 software-layer controls were actually implemented on the real system, since the UL certification covers reference architectures and threat scenarios ETG itself selected, not an automatic guarantee for every possible EtherCAT deployment configuration.
Sources
- EtherCAT Technology Group: EtherCAT — the Ethernet Fieldbus (technology overview, master/slave architecture, processing on the fly)
- EtherCAT Technology Group: Cybersecurity (SL2/IEC 62443-3-3 certification claims, ETG.1800 Security Profile, UL Solutions assessment)
- CISA ICS Advisory ICSA-24-051-02: Industrial Control Systems Network Protocol Parsers (ICSNPP) — Ethercat Zeek Plugin (released 20 February 2024; CVE-2023-7242, -7243, -7244)
- CISA ICS Advisory ICSA-24-312-01: Beckhoff Automation TwinCAT Package Manager (released 7 November 2024)
- CERT@VDE VDE-2024-064: Beckhoff local command injection via TwinCAT Package Manager (CVE-2024-8934)
- Granat, Höfken, Schuba: "Intrusion Detection of the ICS Protocol EtherCAT," 2nd International Conference on Computer, Network Security and Communication Engineering (CNSCE 2017), DEStech Transactions on Computer Science and Engineering
- Kevser Ovaz Akpinar, Ibrahim Ozcelik: "Development of the ECAT Preprocessor with the Trust Communication Approach," Security and Communication Networks, Wiley, 2018
- EtherCAT Technology Group: Built-in Cybersecurity: EtherCAT Delivers Proven Protection (whitepaper, 15 April 2026)
- EtherCAT Technology Group: Cyber Security Requirements on EtherCAT Systems, Explanatory Notes (20 April 2026)
- EtherCAT Technology Group: UL Solutions certifies EtherCAT's Cyber Resilience (press release, 20 April 2026)
- EtherCAT Technology Group: ETG.1800 EtherCAT Security Profile, v0.9.0 draft for review (29 September 2026)
- EtherCAT Technology Group: ETG.1000 EtherCAT Specification series (v1.0.4, 15 September 2017)
- EtherCAT Technology Group: Safety over EtherCAT
- IEC 61158-6-12:2019, Application layer protocol specification, Type 12 elements (Ed. 4.0)
- IEC 61784-3-12:2010+AMD1:2019 CSV, Functional safety fieldbuses, CPF 12 (Ed. 1.1)
- rt-labs: Cybersecurity Advisory for SOEM, RRTL-260921-01 (21 September 2026)
- OpenEtherCATsociety: SOEM v2.0.1 release (15 October 2026)
- NIST NVD: CVE-2026-82944 (SOEM multi-segment transfer overflow)
- CERT@VDE VDE-2020-019: Beckhoff TwinCAT RT network driver EtherLeak (CVE-2020-12494)
- CISA: ICSNPP EtherCAT Zeek plugin repository (cisagov/icsnpp-ethercat)
- NIST Special Publication 800-82 Revision 3, Guide to Operational Technology (OT) Security (September 2023)
About this guide
SoftWSP OT/ICS Desk
Covers the OT/ICS security cluster: 87 guides.

In this cluster
Questions about this page?
