Skip to content

doc(P4sim): add structured example table with ns-3 and P4 links #726

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Mingyumaz wants to merge 2 commits into
base: master
Choose a base branch
from
Open

doc(P4sim): add structured example table with ns-3 and P4 links #726

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
02a17d8
Update P4sim README examples table
Mingyumaz Mar 9, 2026
c6fdf1c
Remove bold in README.md for queuing_test entry
Mingyumaz Mar 16, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
71 changes: 70 additions & 1 deletion doc/P4sim/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -116,6 +116,75 @@ In the [paper](https://dl.acm.org/doi/10.1145/3747204.3747210), P4sim is evaluat

More use cases can be found [here](https://github.com/HapCommSys/p4sim/blob/main/doc/examples.md), demonstrating that P4sim can serve both research and educational purposes, enabling exploration of programmable data-plane behaviors in realistic network contexts.

### High-Performance Simulation with P4sim

**Some of the examples have results and plots for analysis in [link](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/examples_test_result), include the `parameters`, `pcaps` for running, please have a look with more detail.**

| Name | Description | ns-3 script | p4 script |
|-----|-------------|--------------|------|
| [p4-basic-example](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/examples_test_result/p4-basic-example) | [basic](https://github.com/p4lang/tutorials/tree/master/exercises/basic) example | [ns-3](https://github.com/HapCommSys/p4sim/blob/main/examples/p4-basic-example.cc) | basic pipeline verification [p4src](https://github.com/HapCommSys/p4sim/tree/main/examples/p4src/p4_basic) |
| [p4-basic-tunnel](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/examples_test_result/p4-basic-tunnel) | [basic tunnel](https://github.com/p4lang/tutorials/tree/master/exercises/basic_tunnel) example | [ns-3](https://github.com/HapCommSys/p4sim/blob/main/examples/p4-basic-tunnel.cc) | encapsulation / decapsulation test [p4src](https://github.com/HapCommSys/p4sim/tree/main/examples/p4src/basic_tunnel) |
| [p4-fat-tree](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/examples_test_result/p4-fat-tree) | fat tree topo testing | [ns-3](https://github.com/HapCommSys/p4sim/blob/main/examples/topo-fattree.cc) | multi-switch forwarding validation [p4src](https://github.com/HapCommSys/p4sim/tree/main/examples/p4src/fat-tree) |
| [p4-firewall](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/examples_test_result/p4-firewall) | [firewall](https://github.com/p4lang/tutorials/tree/master/exercises/firewall) example | [ns-3](https://github.com/HapCommSys/p4sim/blob/main/examples/p4-firewall.cc) | ACL rule verification [p4src](https://github.com/HapCommSys/p4sim/tree/main/examples/p4src/firewall) |
| [p4-psa-ipv4-forwarding](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/examples_test_result/p4-psa-ipv4-forwarding) | ipv4 forwarding in psa arch | [ns-3](https://github.com/HapCommSys/p4sim/blob/main/examples/p4-psa-ipv4-forwarding.cc) | PSA pipeline example [p4src](https://github.com/HapCommSys/p4sim/tree/main/examples/p4src/simple_psa) |
| [p4-spine-leaf-topo](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/examples_test_result/p4-spine-leaf-topo) | Spine leaf topo testing | [ns-3](https://github.com/HapCommSys/p4sim/blob/main/examples/p4-spine-leaf-topo.cc) | datacenter fabric test [p4src](https://github.com/HapCommSys/p4sim/tree/main/examples/p4src/load_balance) |
| [p4-v1model-ipv4-forwarding](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/examples_test_result/p4-v1model-ipv4-forwarding) | ipv4 forwarding in v1model arch | [ns-3](https://github.com/HapCommSys/p4sim/blob/main/examples/p4-v1model-ipv4-forwarding.cc) | v1model pipeline example [p4src](https://github.com/HapCommSys/p4sim/tree/main/examples/p4src/ipv4_forward) |
| [queuing_test](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/queuing_test) | queuing test with qos priority mapping | [ns-3](https://github.com/HapCommSys/p4sim/blob/main/examples/p4-queue-test.cc) | QoS / priority queue experiment [p4src](https://github.com/HapCommSys/p4sim/tree/main/examples/p4src/qos) |

Following we give two simple examples: `IPv4 Forwarding Benchmark` and `Queue and Packet Scheduling Test` show how run the examples.

#### IPv4 Forwarding Benchmark

The following example runs a simple two-host, one-switch topology with IPv4 forwarding at 100 Mbps. The link rate (`--linkRate`), application data rate (`--appDataRate`), and other parameters can be tuned as needed:

```bash
# V1model architecture (recommended)
./ns3 run p4-v1model-ipv4-forwarding -- \
--pktSize=1000 --appDataRate=100Mbps --linkRate=1000Mbps \
--switchRate=100000 --linkDelay=0.01ms --simDuration=20 --pcap=false

# PSA (Portable Switch Architecture)
./ns3 run p4-psa-ipv4-forwarding -- \
--pktSize=1000 --appDataRate=100Mbps --linkRate=1000Mbps \
--switchRate=100000 --linkDelay=0.01ms --simDuration=20 --pcap=false

# PNA (Portable NIC Architecture) — not yet fully implemented
./ns3 run p4-pna-ipv4-forwarding -- \
--pktSize=1000 --appDataRate=100Mbps --linkRate=1000Mbps \
--switchRate=100000 --linkDelay=0.01ms --simDuration=20 --pcap=false
```

#### Queue and Packet Scheduling Test

To evaluate queuing and packet scheduling behavior on the P4 switch, use the [`p4-queue-test.cc`](https://github.com/HapCommSys/p4sim/blob/main/examples/p4-queue-test.cc) example. It accepts three independent traffic flows with configurable data rates:

```bash
./ns3 run p4-queue-test -- \
--pktSize=1000 \
--appDataRate1=3Mbps --appDataRate2=4Mbps --appDataRate3=5Mbps \
--switchRate=1500 --linkRate=1000Mbps --queueSize=1000 --pcap=true
```

Mote details, results, plots please check [Queue Status Monitor](https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/queuing_test)
Copy link
Contributor

@Dscano Dscano Mar 14, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

According to the Readme https://github.com/HapCommSys/p4sim-artifact-icns3/tree/main/queuing_test we should have the possibilities to play with parameters and also get some csv as output of the test that we can analyze.

I think that would be better specify where p4-queue-test.cc is located and how can be recompiled.

packetA.csv and packet_B.csv are not generated when I'm running

 ./ns3 run p4-queue-test -- \
  --pktSize=1000 \
  --appDataRate1=3Mbps --appDataRate2=4Mbps --appDataRate3=5Mbps \
  --switchRate=1500 --linkRate=1000Mbps --queueSize=1000 --pcap=true

Copy link
Contributor Author

@Mingyumaz Mingyumaz Mar 16, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the feedback. I have updated the README and reorganized the directory to make it clearer.
Running the simulation:

./ns3 run p4-queue-test -- \
  --pktSize=1000 \
  --appDataRate1=3Mbps --appDataRate2=4Mbps --appDataRate3=5Mbps \
  --switchRate=1500 --linkRate=1000Mbps --queueSize=1000 --pcap=true

will generate the *.pcap files.

The packetA.csv and packet_B.csv files are not generated directly by ns-3. They are produced by the script plot_3_0.py, which parses the generated pcap files and extracts packet information for analysis and plotting.

I updated README in the p4sim-artifact-icns3_github now clarifies:

  • where p4-queue-test.cc is located,
  • how to recompile the example,
  • and the workflow (pcap → csv → plot).


> **Note:** Per-port queue parameters cannot currently be set via command-line arguments. Instead, configure them at runtime using the P4 controller command interface:
>
> ```
> set_queue_depth <depth_in_packets> <port_number>
> set_queue_rate <rate_in_pps> <port_number>
> ```
>
> Flow priorities are assigned through match-action table entries that map UDP port numbers to priority levels:
>
> ```
> table_add udp_priority set_priority 2000 => 0x1
> table_add udp_priority set_priority <udp_port_number> => <priority>
> ```
>
> The bottleneck processing rate is controlled by `--switchRate` (in packets per second). In this example it is set to `1500`.

After the simulation completes, inspect the generated PCAP files to observe how packets from the three flows are scheduled and reordered according to their assigned priorities.

## Known Limitations

The packet processing rate `SwitchRate` (in packets per second, pps) must currently be configured manually for each switch. An inappropriate value can cause the switch to enter an idle polling loop, leading to wasted CPU cycles. Automatic rate tuning is planned for a future release.
Expand All @@ -129,4 +198,4 @@ The packet processing rate `SwitchRate` (in packets per second, pps) must curren
**Maintainers & Contributors:**

- **Maintainers**: [Mingyu Ma](mailto:mingyu.ma@tu-dresden.de)
- **Contributors**: Thanks to [GSoC 2025](https://summerofcode.withgoogle.com/) with [Davide](mailto:d.scano89@gmail.com) support and contributor [Vineet](https://github.com/Vineet1101).
- **Contributors**: Thanks to [GSoC 2025](https://summerofcode.withgoogle.com/) with [Davide](mailto:d.scano89@gmail.com) support and contributor [Vineet](https://github.com/Vineet1101).