What Is Multi-Router Looking Glass (MRLG)?
Multi-Router Looking Glass (MRLG) is an open-source network diagnostic tool widely deployed by Internet Service Providers and network operators. It provides a web interface (a "looking glass") that allows network engineers — and sometimes the public — to perform BGP route lookups, traceroutes, and pings from the provider's routers. By querying MRLG, operators can view BGP path information, reachability data, and routing announcements from a given network's perspective.
MRLG consists of a CGI web front-end and a backend process that connects to routers via telnet or SSH to execute diagnostic commands and return results. Because looking glass servers are intentionally internet-facing (their purpose is to provide public network diagnostics), they represent a high-value attack surface: a vulnerability in MRLG directly exposes a host within the network operator's infrastructure to remote exploitation.
Overview
CVE-2014-3931 is a critical buffer overflow vulnerability in Multi-Router Looking Glass (MRLG) that allows an unauthenticated remote attacker to cause arbitrary memory writes and memory corruption, enabling remote code execution on the server hosting the looking glass. With a CVSS score of 9.8 — network-accessible, no authentication required, no user interaction needed — this represents a directly exploitable pre-authentication RCE on internet-facing network infrastructure. CISA added this to the KEV catalog in July 2025, over a decade after the vulnerability was identified, reflecting confirmed exploitation of unpatched looking glass deployments.
Affected Versions
| MRLG Version | Status |
|---|---|
| MRLG versions prior to the fix | Vulnerable |
| Unpatched deployments | Vulnerable |
Many MRLG deployments are on legacy or unmaintained systems where operators have not applied patches. The tool has historically received infrequent updates, and some deployments may be running versions that will never receive vendor patches.
Technical Details
Root Cause: Buffer Overflow in Request Handling
MRLG processes user-supplied input — such as query parameters specifying IP addresses, AS numbers, or hostnames to look up — and passes this data to a backend component that communicates with routers. The buffer overflow occurs when this user input is not properly validated or bounded before being written into a fixed-size stack or heap buffer.
An attacker submitting an oversized or specially crafted input to the MRLG web interface can trigger the overflow, corrupting adjacent memory. Depending on the specific implementation details and system configuration, this can allow:
- Overwriting return addresses on the stack to redirect execution to attacker-controlled code
- Overwriting function pointers or control structures to hijack execution flow
- Arbitrary code execution at the privilege level of the MRLG process (typically the web server user, potentially root on misconfigured systems)
Attack Surface
MRLG deployments are intentionally accessible from the internet — this is their purpose. There is no authentication layer preventing exploitation:
- An attacker can send a crafted HTTP request to the MRLG CGI endpoint
- No credentials or prior access are required
- The vulnerability is exploitable in a single network request
Attack Characteristics
| Attribute | Detail |
|---|---|
| Attack Vector | Network — internet-facing looking glass web interface |
| Authentication | None required |
| User Interaction | None required |
| CVSS Score | 9.8 CRITICAL |
| Impact | Remote code execution on network operator infrastructure |
| CWE | CWE-119: Buffer Overflow |
Discovery
CVE-2014-3931 was identified in 2014 but formally published by NVD in March 2017 — a delayed publication pattern that occurs when vulnerabilities are reported or discovered before the formal CVE process is completed. CISA's addition to the KEV catalog in 2025 confirmed that threat actors were actively exploiting unpatched MRLG deployments more than a decade after the vulnerability was identified.
Exploitation Context
- Internet-facing network infrastructure: Looking glass servers are deployed by ISPs and network operators specifically to be accessible from the internet, meaning no network access bypass is required — exploitation is direct from the public internet
- High-value targets: Compromising a network operator's looking glass server provides a foothold within the operator's network, access to router connectivity and credentials stored in MRLG configuration, and a position for lateral movement into network management infrastructure
- Unpatched legacy deployments: MRLG is a mature tool with infrequent releases; many deployments run old versions that have never been patched, creating a long tail of vulnerable internet-facing servers
- CISA KEV (2025): Added July 2025, confirming active exploitation of unpatched MRLG instances over a decade after the vulnerability was assigned — illustrating the persistent danger of legacy internet-facing network tools
Remediation
-
Apply vendor patch — update MRLG to the latest patched version from the project source. If no patched version is available for your deployment, treat the installation as end-of-life.
-
Decommission if unused — if the looking glass service is no longer actively needed, take it offline. Many organizations maintain historical looking glass servers that are no longer actively monitored or maintained.
-
Restrict access — if the looking glass must remain operational, restrict access to specific trusted IP ranges using firewall rules or web server access controls rather than leaving it openly accessible from the internet.
-
Migrate to maintained alternatives — replace MRLG with a modern, actively maintained looking glass solution with a better security posture and regular updates.
-
Isolate the server — ensure the MRLG host cannot reach sensitive internal network management systems (router management interfaces, TACACS/RADIUS servers, NMS) if it must remain internet-facing, to limit lateral movement in case of compromise.
Key Details
| Property | Value |
|---|---|
| CVE ID | CVE-2014-3931 |
| Vendor / Product | Looking Glass — Multi-Router Looking Glass (MRLG) |
| NVD Published | 2017-03-31 |
| NVD Last Modified | 2025-10-22 |
| CVSS 3.1 Score | 9.8 |
| CVSS 3.1 Vector | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H |
| Severity | CRITICAL |
| CWE | CWE-119 — Improper Restriction of Operations within the Bounds of a Memory Buffer find similar ↗ |
| CISA KEV Added | 2025-07-07 |
| CISA KEV Deadline | 2025-07-28 |
| Known Ransomware Use | No |
CVSS 3.1 Breakdown
Required Action
Timeline
| Date | Event |
|---|---|
| 2014-01-01 | CVE-2014-3931 assigned (vulnerability identified in 2014) |
| 2017-03-31 | CVE-2014-3931 published by NVD (delayed formal publication) |
| 2025-07-07 | Added to CISA Known Exploited Vulnerabilities catalog |
| 2025-07-28 | CISA BOD 22-01 remediation deadline |
References
| Resource | Type |
|---|---|
| NVD — CVE-2014-3931 | Vulnerability Database |
| CISA KEV Catalog Entry | US Government |
| MRLG — Multi-Router Looking Glass | Project |