OpenSSL Patches HollowByte Vulnerability That Enables Denial-of-Service Attacks on TLS Connections with Just 11 Bytes

OpenSSL has released fixes for a vulnerability dubbed HollowByte that allows denial-of-service attacks against servers using unpatched versions of the library. The flaw can be triggered by a malicious TLS request as small as 11 bytes and was addressed without a CVE or formal security advisory because it was classified as a hardening improvement. Researchers from Okta warn that the operational impact can be significant on internet-facing servers, as the issue occurs during the initial TLS handshake when vulnerable OpenSSL versions reserve memory based on the client-declared size before verifying actual data delivery. Attackers can open connections, claim they will send a large message, and then deliver only a tiny portion, leaving the server waiting indefinitely with reserved memory. In tests with NGINX, low-memory servers could be completely taken down while more powerful machines suffered substantial capacity loss without triggering abnormal traffic volumes. The patches are included in OpenSSL versions 4.0.1, 3.6.3, 3.5.7, 3.4.6, and 3.0.21, and administrators are advised to update packages, restart affected services, and monitor for unusual memory growth.

BoletimSecTranslated from Portuguese

Read at source

OpenSSL has issued updates to address a vulnerability known as HollowByte that can be exploited to launch denial-of-service attacks against servers relying on unpatched versions of the widely used cryptographic library.

The flaw allows an attacker to trigger excessive memory reservation during the initial negotiation of TLS connections by sending a specially crafted request of only 11 bytes. Because the issue was treated internally as a hardening measure rather than a critical security problem, it did not receive a CVE identifier or an official security advisory from the OpenSSL project.

Researchers at Okta discovered and reported the weakness, highlighting that its operational impact could be substantial for any server exposed to the internet. The root cause lies in how vulnerable OpenSSL versions allocate memory according to the message size declared by the client before confirming that the promised data will actually arrive.

An attacker can therefore open multiple connections, announce a large incoming message, and then transmit only a minimal portion of data. The server remains in a waiting state, holding the reserved memory until the connection times out or is closed, which can rapidly exhaust available resources.

Testing conducted with the NGINX web server demonstrated that systems with limited memory could be completely disabled, while more powerful machines experienced significant reductions in capacity even though the volume of network traffic did not appear unusual.

Fixed Versions and Recommended Actions

The correction has been incorporated into the following releases: OpenSSL 4.0.1, 3.6.3, 3.5.7, 3.4.6, and 3.0.21.

System administrators are urged to update OpenSSL packages through their distribution’s repositories, restart all services that load the library, and continuously monitor processes exposed to the internet for abnormal memory consumption.

Meridian digestThe week's most consequential security stories, distilled into one briefing.See the digest

More in Vulnerabilities & Exploits

Telegram for Android Patches Voice Call Vulnerability Enabling Code Execution
Yandex Patches Millions of Stations After eMudhra Certificate Keys Compromised
Cisco ASA and FTD Face DoS Vulnerability in SSL VPN Already Exploited in the Wild