feat: initial SyncMmr impl

.github/workflows/ci.yml

@@ -167,11 +167,12 @@ jobs:
           cargo run --bin miden-node-stress-test seed-store \
             --data-directory ${{ env.DATA_DIR }} \
             --num-accounts 500 --public-accounts-percentage 50
-      - name: Benchmark state sync
-        run: |
-          cargo run --bin miden-node-stress-test benchmark-store \
-            --data-directory ${{ env.DATA_DIR }} \
-            --iterations 10 --concurrency 1 sync-state
+  # TODO re-introduce 
+      # - name: Benchmark state sync
+        # run: |
+          # cargo run --bin miden-node-stress-test benchmark-store \
+            # --data-directory ${{ env.DATA_DIR }} \
+            # --iterations 10 --concurrency 1 sync-state
       - name: Benchmark notes sync
         run: |
           cargo run --bin miden-node-stress-test benchmark-store \

CHANGELOG.md

@@ -6,9 +6,13 @@
 
 - [BREAKING] Move block proving from Blocker Producer to the Store ([#1579](https://github.com/0xMiden/miden-node/pull/1579)).
 - [BREAKING] Updated miden-base dependencies to use `next` branch; renamed `NoteInputs` to `NoteStorage`, `.inputs()` to `.storage()`, and database `inputs` column to `storage` ([#1595](https://github.com/0xMiden/miden-node/pull/1595)).
+- [BREAKING] Remove `SynState` and introduce `SyncChainMmr` ([#1591](https://github.com/0xMiden/miden-node/issues/1591)).
+- Introduce `SyncChainMmr` RPC endpoint to sync chain MMR deltas within specified block ranges ([#1591](https://github.com/0xMiden/miden-node/issues/1591)).
 
 ### Changes
 
+- [BREAKING] Removed obsolete `SyncState` RPC endpoint; clients should use `SyncNotes`, `SyncNullifiers`, `SyncAccountVault`, `SyncAccountStorageMaps`, `SyncTransactions`, or `SyncChainMmr` instead ([#1636](https://github.com/0xMiden/miden-node/pull/1636)).
+- Added account ID limits for `SyncTransactions`, `SyncAccountVault`, and `SyncAccountStorageMaps` to `GetLimits` responses ([#1636](https://github.com/0xMiden/miden-node/pull/1636)).
 - [BREAKING] Added typed `GetAccountError` for `GetAccount` endpoint, splitting `BlockNotAvailable` into `UnknownBlock` and `BlockPruned`. `AccountNotFound` and `AccountNotPublic` now return `InvalidArgument` gRPC status instead of `NotFound`; clients should parse the error details discriminant rather than branching on status codes ([#1646](https://github.com/0xMiden/miden-node/pull/1646)).
 - Changed `note_type` field in proto `NoteMetadata` from `uint32` to a `NoteType` enum ([#1594](https://github.com/0xMiden/miden-node/pull/1594)).
 - Refactored NTX Builder startup and introduced `NtxBuilderConfig` with configurable parameters ([#1610](https://github.com/0xMiden/miden-node/pull/1610)).

bin/stress-test/README.md

@@ -20,14 +20,14 @@ This command allows to run stress tests against the Store component. These tests
 
 The endpoints that you can test are:
 - `load_state`
-- `sync_state`
 - `sync_notes`
 - `sync_nullifiers`
 - `sync_transactions`
+- `sync-chain-mmr`
 
 Most benchmarks accept options to control the number of iterations and concurrency level. The `load_state` endpoint is different - it simply measures the one-time startup cost of loading the state from disk.
 
-**Note on Concurrency**: For the endpoints that support it (`sync_state`, `sync_notes`, `sync_nullifiers`), the concurrency parameter controls how many requests are sent in parallel to the store. Since these benchmarks run against a local store (no network overhead), higher concurrency values can help identify bottlenecks in the store's internal processing. The latency measurements exclude network time and represent pure store processing time.
+**Note on Concurrency**: For the endpoints that support it (`sync_notes`, `sync_nullifiers`), the concurrency parameter controls how many requests are sent in parallel to the store. Since these benchmarks run against a local store (no network overhead), higher concurrency values can help identify bottlenecks in the store's internal processing. The latency measurements exclude network time and represent pure store processing time.
 
 Example usage:
 
@@ -119,18 +119,6 @@ Database contains 99961 accounts and 99960 nullifiers
 
 **Performance Note**: The load-state benchmark shows that account tree loading (~21.3s) and nullifier tree loading (~21.5s) are the primary bottlenecks, while MMR loading and database connection are negligible (<3ms each).
 
-- sync-state
-``` bash
-$ miden-node-stress-test benchmark-store --data-directory ./data --iterations 10000 --concurrency 16 sync-state
-
-Average request latency: 1.120061ms
-P50 request latency: 1.106042ms
-P95 request latency: 1.530708ms
-P99 request latency: 1.919209ms
-P99.9 request latency: 5.795125ms
-Average notes per response: 1.3159
-```
-
 - sync-notes
 ``` bash
 $ miden-node-stress-test benchmark-store --data-directory ./data --iterations 10000 --concurrency 16 sync-notes
@@ -171,5 +159,21 @@ Pagination statistics:
   Average pages per run: 2.00
 ```
 
+- sync-chain-mmr
+``` bash
+$ miden-node-stress-test benchmark-store --data-directory ./data --iterations 10000 --concurrency 16 sync-chain-mmr --block-range 1000
+
+Average request latency: 1.021ms
+P50 request latency: 0.981ms
+P95 request latency: 1.412ms
+P99 request latency: 1.822ms
+P99.9 request latency: 3.174ms
+Pagination statistics:
+  Total runs: 10000
+  Runs triggering pagination: 1
+  Pagination rate: 0.01%
+  Average pages per run: 1.00
+```
+
 ## License
 This project is [MIT licensed](../../LICENSE).

bin/stress-test/src/main.rs

@@ -4,9 +4,9 @@ use clap::{Parser, Subcommand};
 use miden_node_utils::logging::OpenTelemetry;
 use seeding::seed_store;
 use store::{
+    bench_sync_chain_mmr,
     bench_sync_notes,
     bench_sync_nullifiers,
-    bench_sync_state,
     bench_sync_transactions,
     load_state,
 };
@@ -70,8 +70,6 @@ pub enum Endpoint {
         #[arg(short, long, value_name = "PREFIXES", default_value = "10")]
         prefixes: usize,
     },
-    #[command(name = "sync-state")]
-    SyncState,
     #[command(name = "sync-notes")]
     SyncNotes,
     #[command(name = "sync-transactions")]
@@ -83,6 +81,12 @@ pub enum Endpoint {
         #[arg(short, long, value_name = "BLOCK_RANGE", default_value = "100")]
         block_range: u32,
     },
+    #[command(name = "sync-chain-mmr")]
+    SyncChainMmr {
+        /// Block range size for each request (number of blocks to query).
+        #[arg(short, long, value_name = "BLOCK_RANGE", default_value = "1000")]
+        block_range: u32,
+    },
     #[command(name = "load-state")]
     LoadState,
 }
@@ -111,9 +115,6 @@ async fn main() {
             Endpoint::SyncNullifiers { prefixes } => {
                 bench_sync_nullifiers(data_directory, iterations, concurrency, prefixes).await;
             },
-            Endpoint::SyncState => {
-                bench_sync_state(data_directory, iterations, concurrency).await;
-            },
             Endpoint::SyncNotes => {
                 bench_sync_notes(data_directory, iterations, concurrency).await;
             },
@@ -127,6 +128,9 @@ async fn main() {
                 )
                 .await;
             },
+            Endpoint::SyncChainMmr { block_range } => {
+                bench_sync_chain_mmr(data_directory, iterations, concurrency, block_range).await;
+            },
             Endpoint::LoadState => {
                 load_state(&data_directory).await;
             },