Milestone: session rekey (~120s rotation) + PQ-hybrid handshake path #9
Labels
No labels
bug
documentation
duplicate
enhancement
good first issue
help wanted
invalid
question
wontfix
No milestone
No project
No assignees
1 participant
Notifications
Due date
No due date set.
Dependencies
No dependencies set
Reference
femboy/yip#9
Loading…
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
The daemon does a single Noise-IK handshake at startup and never rekeys. The design calls for ~120 s session rotation (forward secrecy) and a drop-in path for a Rosenpass-style hybrid PQ handshake.
Scope
SessioninDataPlane, with an overlap window so in-flight packets under the old epoch still decrypt (route by epoch).yip-cryptoalready provides the mechanism (a new handshake = a newSession); the daemon must drive the schedule + epoch handling.pskmodifier onNOISE_PARAMSis reserved but unexercised — wire Classic McEliece + ML-KEM feeding a PSK (Rosenpass model). Likely its own sub-task once classical rekey works.Notes
yip-crypto+bin/yipd(tunnel/dataplane epoch handling). Sequence after io_uring Phase B (#7).Sessionresets counters — the epoch must disambiguate).Rekey half of this milestone is merged to main (#90 = 9a classical ~120s rotation + epoch handling; #92 = #91 relay-path rekey completion). Each peer's Noise-IK session now rotates ~120s with old→new epoch overlap and confirmed-switch, on both direct and relay paths, verified loss-free over ~10 rotations in netns (both drivers).
Remaining: the PQ-hybrid handshake (9b). Design in progress — decided so far: inline (KEM material rides the existing msg1/msg2 payload seam, not a separate Rosenpass-style daemon) + scope A (X25519 + ML-KEM ephemeral → PSK via the reserved
pskNoise modifier, on the 9a rekey path). Classic McEliece static KEM (~524KB pubkey → config/identity distribution) and stateless biscuits deferred to a 9c follow-up. Open sizing question: ML-KEM-512 vs 768 under theOBF_MTU_BUDGET=1200ceiling (512 fits msg1+cert with ~75B headroom; 768 overflows). Paused pending throughput/MTU investigation.