sabe_estimator
Estimate overlay load and safe available bandwidth from existing delay or packet-loss measurements using M/M/1/K equations.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- sabe_estimator
- description
- Estimate overlay load and safe available bandwidth from existing delay or packet-loss measurements using M/M/1/K equations.
02 资源文件 3
- 脚本scripts/sdwan_qos.py6.2 KB
- 参考文档tests/fixture_scenario.json1.3 KB
- 脚本tests/test_sdwan_qos.py1.5 KB
03 SKILL.md 正文
Silent Available Bandwidth Estimator
When To Use
Use this skill when reconstructing or testing the SD-WAN QoS optimization mechanism from the paper. It is appropriate for reduced recovery experiments, validation fixtures, and future implementations that need the paper's silent available bandwidth estimator contract. Do not use it as evidence for full NS3 reproduction by itself.
Inputs
- Scenario JSON with
links,flows, and optionalmeasurements. - Flow records include
id,priority,demand,allowed_links,delay_sla, andloss_sla. - Link records include
idandcapacityin Mbps.
Outputs
- Validated scenario objects, allocation dictionaries, SABE estimates, local-search traces, or SLA metrics depending on the entry point.
- JSON artifacts suitable for
recovery_result.jsonmechanism checks.
Workflow
- Validate the SD-WAN flow and link model before optimization.
- Estimate safe link capacity from passive measurements with the SABE helper when measurements are available.
- Run priority-aware allocation before evaluating SLA satisfaction.
- Preserve trace records showing high-priority reservation and low-priority search increments.
- Report reduced/proxy limitations explicitly when not running packet-level simulation.
Validation
Run python scripts/sdwan_qos.py tests/fixture_scenario.json --output /tmp/sdwan_qos_result.json or python -m pytest tests from the skill directory. The included tests use deterministic fixtures and do not require the original repository.
Limitations
The scripts implement a compact mechanism-faithful proxy rather than the full nonlinear solver or NS3 simulator. Capacity units are Mbps and packet loss is represented as a fraction.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).