OpenSSL 4.1.0 is a feature release that matters because OpenSSL is one of the invisible load-bearing walls of internet infrastructure. When this library adds capability, the entire stack — from embedded IoT devices to cloud load balancers — gains options. When it ships bugs, the entire stack bleeds. This release does both, and understanding the ratio is the point. The headline feature is DTLS 1.3 support per RFC 9147, which brings the security properties of TLS 1.3 — forward secrecy, reduced round trips, encrypted handshake metadata — to datagram protocols like UDP. This matters for real-time applications: VoIP, video conferencing, gaming, IoT telemetry. These sectors have been stuck on DTLS 1.2 while the stream-protocol world moved on. OpenSSL implementing this in its mainline release removes one of the biggest blockers to adoption. The second significant cluster is post-quantum cryptography optimization. ML-DSA (the NIST-standardized digital signature algorithm, formerly CRYSTALS-Dilithium) and ML-KEM (the key encapsulation mechanism, formerly CRYSTALS-Kyber) get optimized NTT (Number Theoretic Transform) operations on both ppc64le and x8664, including AVX-512-accelerated SHAKE x4 operations for ML-DSA. Additionally, AES-CBC decryption gets AVX-512 and VAES optimizations. These are not new algorithms — they're making already-standardized post-quantum algorithms fast enough for production deployment at scale. The difference between 'standardized' and 'deployable' is often raw performance, and this closes that gap. The housekeeping is telling. Dropping Windows-on-Itanium (VC-WIN64I) and Windows CE targets removes dead weight that consumed maintenance cycles. Adding MSVC 2013 bridge targets for C99 gaps is pragmatic — legacy enterprise environments still run on old toolchains, and refusing to support them doesn't make them disappear. The tsget utility's migration from the abandoned WWW::Curl::Easy to Net::Curl::Easy is minor but signals active dependency hygiene. The known issues section deserves attention. Four open DTLS 1.3 bugs shipped with the release, all involving edge cases in retransmission, error reporting, and handshake completion. Issue #33242 — where servers abort handshakes when clients acknowledge flight messages out of expected order — is the kind of interoperability landmine that can delay real-world DTLS 1.3 adoption. These are not showstoppers, but they reveal that DTLS 1.3 support is early-stage, likely to burn early adopters in production. The release also adds RFC 8701 GREASE support (injecting random unknown values into handshakes to keep implementations from ossifying around known-good values), IKEv2 KDF support (useful for VPN implementations), DTLS listener API enhancements, and initial Elbrus2000 (e2k) architecture support — a Russian-designed processor family. The e2k addition is architecturally minor but geopolitically notable: it signals OpenSSL's continued posture as a universal infrastructure library, not one that picks sides on hardware ecosystems. Net assessment: this is a genuinely generative release. It expands what the internet's most widely deployed crypto library can do, makes post-quantum algorithms faster, cleans up legacy debt, and ships an important new protocol implementation — while being honest about known defects. The DTLS 1.3 bugs are the price of shipping early rather than waiting for perfection, a tradeoff that favors the ecosystem getting real-world testing data.
