Skip to content

feat(gms): add host-tier storage contracts [GMS 06/18]#11051

Closed
hutm wants to merge 1 commit into
review/gms-v2-latehost-04-sglang-integrationfrom
review/gms-v2-latehost-08-1-host-tier-offload-common
Closed

feat(gms): add host-tier storage contracts [GMS 06/18]#11051
hutm wants to merge 1 commit into
review/gms-v2-latehost-04-sglang-integrationfrom
review/gms-v2-latehost-08-1-host-tier-offload-common

Conversation

@hutm

@hutm hutm commented Jun 29, 2026

Copy link
Copy Markdown
Contributor

Summary

Define the engine-neutral contracts for moving KV content between GPU memory, host memory, and durable storage.

Scope

Adds ring, memory, storage, and transfer interfaces only. Daemon lifecycle, engine adapters, and optional network backends remain separate.

Stack

Position 6/18 in the GMS-managed KV review train. This PR is based on review/gms-v2-latehost-04-sglang-integration.

Parent: #10450
Tracking: #10455

Review migration

This is the current review entry replacing historical merged PR #10472. GitHub does not allow a merged PR to be retargeted; #10472 and all of its comments and reviews remain intact as the historical record.

The train is rebased on main at aeda23b483831df2f75b697b8ccf65f4ffd9f3d6. The reordered functionality is preserved while each PR boundary is independently buildable and lintable: compatibility call sites and the engine-facing placement adapter now appear at first use; missing type imports and test markers were added; repository-pinned formatting was applied; one unused constant was removed; and every commit carries a DCO sign-off. The complete range passes git diff --check and the full pre-commit suite.

Validation

Contract and component tests are included with the host-tier implementation; cross-node behavior is completed in positions 8-9.

@hutm hutm requested review from a team as code owners June 29, 2026 17:34
@hutm hutm requested a review from a team June 29, 2026 17:34
@copy-pr-bot

copy-pr-bot Bot commented Jun 29, 2026

Copy link
Copy Markdown

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@datadog-official

This comment has been minimized.

@hutm hutm force-pushed the review/gms-v2-latehost-08-1-host-tier-offload-common branch from 06ad30d to a202b9b Compare June 29, 2026 17:45
@hutm hutm force-pushed the review/gms-v2-latehost-04-sglang-integration branch from bac3fdb to 47f9a1a Compare June 29, 2026 17:45
@hutm hutm changed the title [GMS 06/18] Host-tier ring, memory, and storage contracts feat(gms): add host-tier storage contracts [GMS 06/18] Jun 29, 2026
@github-actions github-actions Bot added the feat label Jun 29, 2026
@hutm hutm force-pushed the review/gms-v2-latehost-04-sglang-integration branch from 47f9a1a to ed08b35 Compare June 29, 2026 17:57
@hutm hutm force-pushed the review/gms-v2-latehost-08-1-host-tier-offload-common branch 2 times, most recently from c5ed3d0 to 403de8b Compare July 5, 2026 19:42
@hutm hutm force-pushed the review/gms-v2-latehost-04-sglang-integration branch 2 times, most recently from 275d776 to 67dd9d9 Compare July 8, 2026 23:40
@hutm hutm force-pushed the review/gms-v2-latehost-08-1-host-tier-offload-common branch from 403de8b to 0ffff30 Compare July 8, 2026 23:40
@hutm hutm force-pushed the review/gms-v2-latehost-04-sglang-integration branch from 67dd9d9 to 42c69ca Compare July 8, 2026 23:47
@hutm hutm force-pushed the review/gms-v2-latehost-08-1-host-tier-offload-common branch 3 times, most recently from c3b6f4c to 7d871e8 Compare July 9, 2026 00:13
@hutm hutm force-pushed the review/gms-v2-latehost-04-sglang-integration branch from 42c69ca to 3ece523 Compare July 9, 2026 00:13
Signed-off-by: mkhadkevich <mkhadkevich@nvidia.com>
@hutm hutm force-pushed the review/gms-v2-latehost-08-1-host-tier-offload-common branch from 7d871e8 to de02314 Compare July 9, 2026 00:27
@hutm

hutm commented Jul 13, 2026

Copy link
Copy Markdown
Contributor Author

Closing this PR because the deprecated umbrella #10450 has been replaced by the deliberately vLLM-only MVP tracker #11600. This PR covers work outside that MVP and is deferred to a future milestone; it is not being marked implemented, rejected, or unnecessary. The branch, comments, and reviews are intentionally retained so future work can reference the design and discussion.

@hutm hutm closed this Jul 13, 2026
@hutm hutm deleted the review/gms-v2-latehost-08-1-host-tier-offload-common branch July 13, 2026 14:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant