Commit graph

13 commits

Author SHA1 Message Date
ce57012e11 feat: rework datastream into catalog events, add gguf metadata
- Rework the datastream endpoint into a catalog/event model (DatastreamEvent,
  SubscriptionRequest, channel/stream descriptors, DatastreamPublisherActor) across
  endpoint/frame/mux/wire/views.
- Add gguf_metadata planning reader, a dashboard hardware view, and iroh-driver
  datastream transport.
- Grow mvp-system orchestrator/worker_node bins and staging/provisioning; rename
  mvp_one_node_chat->mvp_chat; extend tinygrad worker.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-07-12 10:14:34 +04:00
6f1c048669 refactor: retire old pipeline-parallel app, restructure mvp-system
- Drop the standalone apps/old-pipeline-parallel-inference app and its
  fleet/orchestrator/tests (~24k lines).
- Add datastream::hardware (cpu/gpu/net) modules; add mvp-system config, arena_manager,
  and vastai_offer_preview.
- Rename mvp_orch_one_node->orchestrator and mvp_node->worker_node; expand VastAI
  provisioning; rework xtask runner.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-07-07 14:40:02 +04:00
a97864f7ec feat(mvp-system): expand one-node mvp binaries and provisioning
Flesh out mvp_node, mvp_one_node_chat, and mvp_orch_one_node binaries. Add node_image
and relay_provisioning; grow provisioning and vastai. Tune datastream mux/timing/emit.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-07-05 13:59:51 +04:00
0ffc5fd275 feat: add mvp provisioning subsystem and datastream transport
mvp-system: provisioner actor, provisioning module, node_agent, dashboard_view,
observability_surface; expand gpu_worker ctl/ingress/egress and run_plan.
iroh-driver: replace relay binary with datastream_transport; datastream gains endpoint
abstraction. Archive pipeline-parallel-inference app to old-pipeline-parallel-inference.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-06-25 16:30:18 +04:00
f4dee176c0 refactor: extract iroh-driver crate, drop node example
Move iroh_driver and the relay binary out of distribution into a dedicated
crates/iroh-driver (lib re-exports IrohDriver; relay bin renamed). Remove the node crate
and the single-gpu-inference example; drop the docker/datastream demo. Slim
pipeline-parallel vastai.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-06-24 00:10:41 +04:00
0d1b95695d refactor: drop datastore crate, stale specs, and benches
Remove the datastore crate, the top-level design/orchestration/ring specs, the
benches, and the ci config. Add the dashboard host telemetry sampler
(cpu/disk/net/gpu/mem). Localize the pipeline-parallel e2e stub/mock paths.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-06-23 19:42:28 +04:00
1f8a67231c feat(sim): rebuild around discrete-event engine
Replaces the generic SimNode/gossip/dashboard framework with a virtual-time
discrete-event engine (priority queue ordered by time/node/fiber/seq), a TOML spec
parser, bundle writer, replay, divergence detector, lint, and postproc, plus the
parity-bar test harness with fixtures and xtask parity-lock tooling. Rewrites
transport identity/crypto and adds the SPEC/TESTING_SPEC/OBSERVABILITY/NORTH_STAR docs.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-05-21 23:48:02 +04:00
19fabb707e refactor: consolidate crate functions (#50)
Remove co-dependencies for different modules found in `crates` and migrate the development history to a new repository. The docs were stale, and largely not getting used, so simply deleted for now. When code stabilizes more, they will become useful again.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-02-24 09:12:28 +00:00
33a217c96a feat: begin formal verification pipeline (#49)
Initial experiments in 'formal verification' of authorization tooling. Current state is not to be trusted, but we are not in a security critical situation, so that is fine.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-02-23 04:53:04 +00:00
599678e7c9 feat: cluster test improvements; datastore pools (#45)
Introduce a gossip-converged pooled datastore protocol layered on SWIM piggybacking, backed by a reusable gossip-channel abstraction, plus a Docker-free multi-process cluster test runner.

- crates/distribution/src/gossip_channel.rs: add the GossipChannel trait (piggyback on SWIM messages) and a budget-limited DisseminationBuffer<T> that replaces the four duplicated Lambda*ceil(log2(n)) dissemination copies
- crates/datastore/src/pool: add PoolDisseminator (CRDT state for membership/capacity/content-location/ACL with join/leave/announce) and PoolCoordinator (placement-aware CRUD actor delegating to the co-located DatastoreNode)
- crates/shared-types/src/pool.rs: add shared pool protocol types (PoolId plus member/capacity/content-location/ACL entries and PoolConfig) consumed by both distribution and datastore
- crates/dashboard/src/pool_html.rs: add a live pool dashboard page (membership, capacity, content locations) and add pool_tests integration coverage
- xtask/src/sim_cluster.rs: add the sim-cluster runner that spawns N swactor nodes over iroh through a local relay server, reusing the docker cluster scenarios without Docker
- crates/distribution/src/iroh_driver.rs: add relay-URL resolution (cache, then SWIM gossip, then home relay) with a 2s connect timeout to back the relay-based connections

Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-02-20 17:30:37 +00:00
c6cb88e335 feat: stability for deployment and distribution (#44)
Make distribution and deployment more stable. Consolidate the logic for a generic swactor node.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-02-19 14:39:33 +00:00
9f4ccdc151 feat: MVP authorization layer for swactor datastores (#43)
Very barebones, untrustworthy, barely reviewed auth layer. LGTM.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-02-16 15:39:27 +00:00
8412d01393 feat: content addressed datastore (#41)
Content addressable datastore. Allows you to configure a node to store and stream large blobs of data, and retrieve them from any swactor-connected node.


Signed-off-by: Zachery Aaron Shores-Chmielewski <zacheryasc@gmail.com>
2026-02-15 17:03:31 +00:00