docs: update benchmark references after experimental benchmarks removal

- Update transport_perf_analysis.md to note experimental benchmarks were removed - Rename "blackbox" references to "transport" in transport_benchmark_methodology.md - Fix benchmark command examples to use current group names 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

Author: iduartgomez

docs/architecture/transport_benchmark_methodology.md

@@ -236,20 +236,20 @@ fn bench_serialize_pure(c: &mut Criterion) {
 
 Uses channels instead of sockets - measures protocol logic without syscalls.
 
-**Running the Blackbox Benchmarks:**
+**Running the Transport Benchmarks:**
 
 ```bash
-# Run just the blackbox transport benchmarks (uses actual transport code with mock I/O)
-cargo bench --bench transport_perf --features bench -- blackbox
+# Run just the transport benchmarks (uses actual transport code with mock I/O)
+cargo bench --bench transport_perf --features bench -- transport
 ```
 
-The `blackbox` benchmark group tests the **actual transport code** with mock sockets:
+The `transport` benchmark group tests the **actual transport code** with mock sockets:
 
 | Benchmark | What it measures |
 |-----------|-----------------|
-| `level1_blackbox/connection/establish` | Full handshake: key exchange, encryption setup |
-| `level1_blackbox/throughput/bytes/*` | End-to-end message throughput at different sizes |
-| `level1_blackbox/fast_channel/*` | Our crossbeam-based channel vs tokio::sync::mpsc |
+| `transport/connection/establish` | Full handshake: key exchange, encryption setup |
+| `transport/throughput/bytes/*` | End-to-end message throughput at different sizes |
+| `transport/fast_channel/*` | Our crossbeam-based channel vs tokio::sync::mpsc |
 
 **Using the Mock Transport Infrastructure:**
 
@@ -550,12 +550,12 @@ echo "=== Results in target/criterion/ ==="
 | Encryption throughput | `level0` | Any | None | Perfect |
 | Serialization | `level0` | Any | None | Perfect |
 | Protocol state machines | `level0` | Any | None | Perfect |
-| **Blackbox transport** | `blackbox` | Any | **Channels only** | **High** |
+| **Transport (mock I/O)** | `transport` | Any | **Channels only** | **High** |
 | Mock I/O throughput | `level1` | Linux bare metal | Channels only | High |
 | Syscall overhead | `level2` | Linux bare metal | recv/send | Medium |
 | Full stack | `level3` | Controlled hardware | Everything | Low |
 
-**Key insight**: The `blackbox` benchmark group tests the **actual Freenet transport code** (PeerConnection, fast_channel, encryption) with mock I/O. This catches real regressions in the transport pipeline without kernel noise.
+**Key insight**: The `transport` benchmark group tests the **actual Freenet transport code** (PeerConnection, fast_channel, encryption) with mock I/O. This catches real regressions in the transport pipeline without kernel noise.
 
 To run all benchmark groups:
 ```bash
@@ -565,7 +565,7 @@ cargo bench --bench transport_perf --features bench
 To run specific groups:
 ```bash
 cargo bench --bench transport_perf --features bench -- level0 # Pure logic
-cargo bench --bench transport_perf --features bench -- blackbox  # Actual transport with mock I/O
+cargo bench --bench transport_perf --features bench -- transport # Actual transport with mock I/O
 cargo bench --bench transport_perf --features bench -- level2 # Loopback
 ```
 

docs/architecture/transport_perf_analysis.md

@@ -636,20 +636,16 @@ criterion_main!(benches);
 
 ### 11.4 Running the Benchmarks
 
-```bash
-# Run packet size experiments
-./scripts/run_benchmarks.sh experimental_packet
-
-# Run Tokio overhead experiments
-./scripts/run_benchmarks.sh experimental_tokio
+> **Note**: The experimental benchmarks referenced in sections 11.1-11.3 were one-time experiments
+> to inform design decisions. They have been removed from the codebase. The results are preserved
+> here for reference. To run current benchmarks:
 
-# Run full pipeline experiments
-./scripts/run_benchmarks.sh experimental_combined
+```bash
+# Run transport layer benchmarks (actual code with mock I/O - used in CI)
+cargo bench --bench transport_perf --features bench -- transport
 
-# Or run directly with cargo (requires bench feature):
-cargo bench --bench transport_perf --features bench -- experimental_packet
-cargo bench --bench transport_perf --features bench -- experimental_tokio
-cargo bench --bench transport_perf --features bench -- experimental_combined
+# Run all benchmark levels
+cargo bench --bench transport_perf --features bench
 ```
 
 ---
@@ -761,22 +757,19 @@ This means:
 
 ### 12.7 Running These Benchmarks
 
-```bash
-# Run full size×batch matrix
-./scripts/run_benchmarks.sh experimental_size_batch
+> **Note**: The size×batch experimental benchmarks referenced in section 12 were one-time
+> experiments to inform design decisions. They have been removed from the codebase.
+> The results are preserved here for reference.
 
-# Individual benchmark groups (requires bench feature)
-cargo bench --bench transport_perf --features bench -- experimental_size_batch_matrix
-cargo bench --bench transport_perf --features bench -- experimental_batch_improvement
-cargo bench --bench transport_perf --features bench -- experimental_throughput_ceiling
+```bash
+# Run transport layer benchmarks (actual transport code with mock I/O)
+cargo bench --bench transport_perf --features bench -- transport
 
-# NEW: Blackbox transport benchmarks (actual transport code with mock I/O)
-cargo bench --bench transport_perf --features bench -- blackbox
+# Run all benchmark levels (level0, level1, level2, level3, transport)
+cargo bench --bench transport_perf --features bench
 ```
 
 ---
 
 *Analysis performed: December 2025*
-*Experimental benchmarks added: December 2025*
-*Size×batch interaction analysis: December 2025*
 *Codebase version: commit c50d888*