dperf <a href="https://hellogithub.com/repository/67958cc5d1f44a6a84f3544e3c007e5f" target="_blank"><img src="https://abroad.hellogithub.com/v1/widgets/recommend.svg?rid=67958cc5d1f44a6a84f3544e3c007e5f&claim_uid=Thc9mJByaKSbdng&theme=small" alt="Featured｜HelloGitHub" /></a>

dperf is a high-performance network traffic generator and load testing tool based on DPDK.

Advantages

• High Performance

  Built on DPDK, dperf can generate massive traffic using a single x86 server — achieving tens of millions of HTTP Connections Per Second (CPS), hundreds of Gbps throughput, and billions of concurrent connections.

• Comprehensive Statistics

  Provides detailed real-time metrics and identifies every packet drop or error.

• Versatile Use Cases

  • Load and stability testing for Layer 4 Load Balancers and gateways
  • Network performance benchmarking for cloud servers
  • Evaluation of NIC and CPU packet processing capabilities
  • Acts as a high-performance HTTP server or client for testing scenarios

Performance

All results were measured using the following hardware and configuration:

• CPU: Intel Xeon Gold 5418Y × 2
• Memory: 128 GB × 2
• NIC: Intel E810-C Dual-Port 100GbE × 2
• OS: Linux 5.10.0

HTTP Connections Per Second (CPS)

| Client Cores | Server Cores | HTTP CPS (Million) | Client CPU Usage (%) | Server CPU Usage (%) | |--------------|---------------|-------------------|----------------------|----------------------| | 1 | 1 | 4 | 74 | 71 | | 2 | 2 | 8 | 74 | 72 | | 4 | 4 | 16 | 73 | 70 | | 8 | 8 | 32 | 70 | 68 | | 16 | 16 | 64 | 70 | 68 |

HTTP Throughput (Gbps)

| Client Cores | Server Cores | RX Throughput | TX Throughput | Client CPU Usage (%) | Server CPU Usage (%) | |--------------|--------------|---------------|---------------|----------------------|----------------------| | 1 | 1 | 98.3 Gbps | 98.3 Gbps | 78 | 80 | | 2 | 2 | 196.7 Gbps | 196.7 Gbps | 78 | 82 |

Concurrent Connections

| Client Cores | Server Cores | Connections (Billion) | Client CPU Usage (%) | Server CPU Usage (%) | Memory Usage (GB) | |--------------|--------------|-----------------------|----------------------|----------------------|-------------------| | 1 | 1 | 1 | 48 | 48 | 60 | | 2 | 2 | 2 | 48 | 48 | 120 | | 4 | 4 | 4 | 48 | 48 | 240 |

TCP Packets Per Second (PPS)

| Client Cores | RX PPS (Mpps) | TX PPS (Mpps) | Client CPU Usage (%) | |--------------|---------------|---------------|----------------------| | 1 | 16.8 | 16.8 | 99 | | 6 | 105.2 | 105.2 | 99 | | 12 | 204.6 | 204.6 | 99 |

Real-Time Statistics

dperf prints real-time statistics every second, including CPS, TPS, PPS, packet drops, socket errors, and HTTP status counts. Example output:

seconds 22                 cpuUsage 52
pktRx   3,001,058          pktTx    3,001,025          bitsRx   2,272,799,040      bitsTx  1,920,657,600      dropTx  0
arpRx   0                  arpTx    0                  icmpRx   0                  icmpTx  0                  otherRx 0          badRx 0
synRx   1,000,345          synTx    1,000,330          finRx    1,000,350          finTx   1,000,350          rstRx   0          rstTx 0
synRt   0                  finRt    0                  ackRt    0                  pushRt  0                  tcpDrop 0
skOpen  1,000,330          skClose  1,000,363          skCon    230                skErr   0
httpGet 1,000,345          http2XX  1,000,350          httpErr  0
ierrors 0                  oerrors  0                  imissed  0

Documentation

See the official website at https://dperf.org/.

Contributing

We welcome contributions! Please see the CONTRIBUTING file for details.

Patent

• Title: Testing Method and Apparatus for Network Devices
• Inventor: Jianzhang Peng
• Patent Number: CN114205274B
• Issue Date: June 11, 2024

Acknowledgment

We gratefully acknowledge xnetin for providing the high-performance testing platform used in our benchmarking experiments.

Author

Jianzhang Peng holds a Ph.D. in Computer Science from the University of Science and Technology of China (USTC). He previously worked as a Principal Engineer at Baidu, where he contributed to the development of high-performance L4 load balancer systems. He initiated and developed the dperf project during his time at Baidu, and continues to maintain it as an open-source contributor. His current focus is on low-latency network protocol stacks for quantitative trading systems.

License

dperf is licensed under the Apache License 2.0.