CVE-2021-38447 is an asymmetric resource consumption flaw (CWE-405) in OCI OpenDDS before 3.18.1: a malformed RTPS payload makes the allocator grab 1-byte chunks in a loop, so one packet can stall a DDS node or ROS 2 robot. NVD scores it 7.5, CISA 8.6. OCI fixed it in OpenDDS 3.18.1.
Vulnerability summary
CVSS base score
7.5
High severity
EPSS score (FIRST)
2.1%
Probability of exploitation in the wild in the next 30 days
CISA KEV catalog
No
Not listed
Published
Vulnerability ID: CVE-2021-38447 · Source: NVD
Affected components
| Component | CPE | Version range |
|---|---|---|
| OCI OpenDDS | cpe:2.3:a:objectcomputing:opendds:*:*:*:*:*:*:*:* | before 3.18.1 (NVD, CISA ICSA-21-315-02); fixed in 3.18.1 |
What does this mean for affected devices?
NVD's description: "OCI OpenDDS versions prior to 3.18.1 are vulnerable when an attacker sends a specially crafted packet to flood target devices with unwanted traffic, which may result in a denial-of-service condition." NVD's Primary score is 7.5, CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H, availability only. CISA, the assigning CNA, scores it higher at 8.6 (AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:H) and classes it CWE-405, asymmetric resource consumption (NVD).
What does the malformed packet do?
The NVD text reads like the amplification bug CVE-2021-38429, but the researchers describe a different mechanism. Their white paper files CVE-2021-38447 under the "malformed RTPS packet" attack surface and calls it "Slowloris behavior by forcing the allocator to allocate 1-byte chunks in a loop via malformed RTPS payload". They used it with CVE-2021-38445 against a simulated ROS 2 autonomous-driving robot that runs the same software stack as a TurtleBot3: both flaws come from OpenDDS not handling the length of the PID_BUILTIN_ENDPOINT_QOS parameter in an RTPS DATA submessage, and setting parameterLength to 4 null bytes was enough to build the exploit payload (Trend Micro Research).
Which version fixes it?
CISA's ICS advisory ICSA-21-315-02, first published on 2021-11-11 and updated on 2022-02-01, says: "OCI recommends users update to Version 3.18.1 of OpenDDS or later." The researchers list it as patched from 3.18.1. The OpenDDS 3.18.1 release notes, dated 2021-09-24, do not name the CVE but list "Input validity check for RTPS Parameters (#2917), sequences (#2944), addresses (#2981, #3016), locators (#3000)".
What should device makers do?
OpenDDS is often built from source into a product, so the version string may not survive in the binary. Check the build manifest or firmware SBOM for 3.18.1 or later. A node that stalls on one packet is a safety issue on a robot or vehicle, so also keep RTPS traffic inside the control zone and away from untrusted networks, as CISA advises. Our DDS security guide covers segmentation, and the input-handling fix guide covers resource-exhaustion weaknesses.
CVE-2021-38447 is not in the CISA KEV catalog as of 2026-11-18 (catalog version 2026.11.17). FIRST EPSS gives it 0.0214, the 82.3rd percentile, in the 2026-11-17 dataset (CISA KEV, FIRST EPSS).
Evidence
- OpenDDS before 3.18.1, crafted packet floods targets, DoS; NVD Primary 7.5 AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H; CISA 8.6 AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:H, CWE-405; CPE opendds before 3.18.1; published 2022-05-05. — NVD
- CVE-2021-38447 CWE-405 CVSS v3 8.6; OpenDDS all versions prior to 3.18.1; "OCI recommends users update to Version 3.18.1 of OpenDDS or later"; ICSA-21-315-02 initial release 2021-11-11, Update A 2022-02-01; sectors "Multiple", deployed "Worldwide"; general measures: minimize network exposure, firewalls, VPN. — CISA CSAF ICSA-21-315-02
- Table 6: CVE-2021-38447 OpenDDS, ROS 2, patched >= 3.18.1, Slowloris behavior by forcing the allocator to allocate 1-byte chunks in a loop via malformed RTPS payload; Table 10 and section on simulated autonomous-driving robot (same stack as TurtleBot3); CVE-2021-38447 and 38445 from PID_BUILTIN_ENDPOINT_QOS length handling in RTPS DATA submessage; parameterLength set to 4 null bytes. — Trend Micro Research white paper
- OpenDDS 3.18.1 released 2021-09-24; input validity checks for RTPS parameters (#2917), sequences (#2944), addresses (#2981, #3016), locators (#3000); CVEs not named. — OpenDDS on GitHub
- Not in the CISA KEV catalog, version 2026.11.17 (1,782 entries), checked 2026-11-18. — CISA KEV
- EPSS 0.02137, percentile 0.8233, dataset 2026-11-17, read 2026-11-18. — FIRST EPSS

Related advisories
Sources
- NVD
- CISA ICS Advisory ICSA-21-315-02: Multiple Data Distribution Service (DDS) Implementations
- CISA CSAF record for ICSA-21-315-02 (revision 2, 2022-02-01)
- OpenDDS 3.18.1 release notes
- Trend Micro Research white paper: A Security Analysis of the Data Distribution Service (DDS) Protocol
- Trend Micro Research: Critically Underrated: Studying the Data Distribution Service (DDS) Protocol
- CISA KEV
- FIRST EPSS
Frequently asked questions
Which OpenDDS version fixes CVE-2021-38447?
OpenDDS 3.18.1 or later. CISA quotes OCI recommending 3.18.1 or later, and the researchers list the flaw as patched from 3.18.1 (per CISA ICSA-21-315-02 and the Trend Micro white paper).
Is CVE-2021-38447 the same as the OpenDDS amplification bug CVE-2021-38429?
No. NVD uses similar wording for both, but CVE-2021-38429 is RTPS discovery amplification (CWE-406), while CVE-2021-38447 is asymmetric resource consumption (CWE-405) where a malformed payload makes the allocator loop on 1-byte chunks (per CISA and the Trend Micro white paper).
Why do NVD and CISA scores differ for CVE-2021-38447?
NVD scores availability impact only (7.5). CISA adds low confidentiality and integrity impact (8.6). Both use a network vector with no privileges or user interaction (per NVD and CISA).
Is CVE-2021-38447 exploited in the wild?
It is not in the CISA KEV catalog as of 2026-11-18 (catalog version 2026.11.17), and CISA's advisory reported no known public exploits targeting these DDS flaws. FIRST EPSS gives it 0.0214, the 82.3rd percentile, in the 2026-11-17 dataset (per CISA KEV, CISA ICSA-21-315-02 and FIRST EPSS).
Questions about this page?