Skip to content

docs: design hierarchical external KV cache - #1

Merged
feichai0017 merged 1 commit into
mainfrom
docs/hierarchical-kv-cache
Nov 15, 2026
Merged

feichai0017 merged 1 commit into
mainfrom
docs/hierarchical-kv-cache

Conversation

@feichai0017

Copy link
Copy Markdown
Contributor

This proposal extends the existing prefix-storage owners into a hierarchical external KV interface. OrbitInfer keeps scheduling, prefix matching and device-page leases; a selected provider owns DRAM/SSD/peer residency. The first integration is OrbitKV, followed by separate Mooncake Store and LMCache adapters.

The design separates logical outcome, borrowed-buffer retirement and cache publication; requires representation identity and bounded byte/operation accounting; and retains the existing lookup/load/store boundary. It includes failure controls and phased implementation. No adapter or runtime behavior is implemented here.

The parallelism guide now distinguishes imported TP/attention-DP/CP/EP/expert-TP source from unimplemented PP and encoder/prefill/decode disaggregation, and from installed multi-GPU qualification.

Validation on the local Apple Silicon development environment:

  • uv run --no-sync python scripts/check.py: passed (format, Clippy, 310 Rust tests, doctests, Ruff, 2,405 Python tests passed / 551 skipped, installed CPU example, core and model wheel boundaries).
  • python3 scripts/check_source.py: all 616 imported files and provenance retained.
  • Local documentation links/anchors: 72 passed; git diff --check: passed.

The first main CI run already has unrelated failures at the base commit: new stable Clippy diagnostics, Linux missing tilelang during collection, and macOS missing torchvision/model package import in child processes. Apple Silicon tests and lint/types passed. The plan records clean-environment CI as a prerequisite to provider implementation; this documentation-only PR does not fix or conceal those failures.

Handoff:

  • Commit: 10ef0ee; baseline: 200cac4.
  • Scope: design for review, not an accepted implementation.
  • Changes: six documentation files; +335 / −11 lines, net +324. No source, tests, dependencies or workflows changed.
  • References: existing Rust/Python storage owners and runner capabilities; OrbitKV architecture read-only at 01d9e36c; official Mooncake Store / LMCache interface documentation checked 2026-11-15.
  • Remaining work: review the proposal, repair the import CI baseline, prove actual-backing retention (P0), then implement the provider sequence.
  • No OrbitKV/OrbiTE changes, GPU jobs, package publication or model downloads.

Copilot AI balanced review requested due to automatic review settings November 15, 2026 04:43

Copilot AI left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota. Try requesting a review at a lower effort level.

An organization owner can choose the repository organization as the billing source for reviews requested by members with a paid Copilot license. Learn about Copilot code review billing.

@feichai0017
feichai0017 merged commit a1c55dd into main Nov 15, 2026
2 of 5 checks passed
@feichai0017
feichai0017 deleted the docs/hierarchical-kv-cache branch November 15, 2026 11:46
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants