What are the main features of baidu/dperf?

The main features of baidu/dperf are: Benchmarking Tools, Load Testing Tools, Zero-Copy Packet Processing, Network Throughput Measurement Tools, Line-Rate Packet Senders, Packet Processing Speed Benchmarks, Stateless TCP Simulation, User-Space Implementations.

Which projects share features with baidu/dperf?

Projects with overlapping indexed features include: microsoft/ethr — ethr is a command line network diagnostic tool and performance analyzer designed to measure throughput, latency, and… codesenberg/bombardier — Bombardier is a concurrent HTTP load generator and performance benchmarking tool written in Go. It functions as a… firehol/netdata — Netdata is a real-time infrastructure monitoring tool and multi-node observability platform. It functions as a… mcollina/autocannon — Autocannon is a Node.js HTTP load tester and benchmarking utility used to measure server throughput and latency. It… libnyanpasu/clash-nyanpasu — Clash Nyanpasu is a network proxy client and graphical user interface for managing Clash network cores. It functions… link1st/go-stress-testing — This project is a performance benchmarking framework and network protocol stress tester designed to evaluate system…

Dperf - generate massive network traffic

Name: baidu/dperf
Author: baidu

dperf is a high-performance network testing tool that generates massive multi-protocol traffic for benchmarking network devices and servers. It uses DPDK to bypass the kernel network stack, enabling line-rate packet generation and capture with zero-copy processing, and implements a full TCP/IP stack in user space to achieve millions of connections per second from a single server.

The tool is designed for stress-testing Layer 4 load balancers and gateways, capable of simulating billions of concurrent TCP or UDP sessions and generating tens of millions of new connections per second. It supports symmetric RSS traffic distribution for accurate load spreading across NIC queues, and uses stateless connection simulation to maintain enormous session counts without full socket state overhead. Real-time per-second metrics including connection rates, throughput, packet drops, and HTTP status counts are collected and printed using lock-free ring buffers.

dperf operates across x86-64, ARM64, and LoongArch processors, and can be deployed inside Docker containers for simplified orchestration in CI or lab environments. It generates mixed protocol traffic covering HTTP, TCP, UDP, IPv4/IPv6, VLAN, and VXLAN, and includes packet mirroring via DPDK pdump for offline analysis. The tool also measures per-connection latency and reports actual CPU utilization alongside traffic metrics.

Features

Benchmarking Tools - Evaluates the connection capacity and stability of Layer 4 load balancers and gateways under billions of concurrent sessions.
Load Testing Tools - A high-performance traffic generator that uses DPDK to create massive HTTP, TCP, and UDP loads for benchmarking.
Zero-Copy Packet Processing - Bypasses the kernel network stack using DPDK for zero-copy packet processing and line-rate generation.
Network Throughput Measurement Tools - Saturates network throughput by generating high-bandwidth traffic streams to measure maximum throughput.
Line-Rate Packet Senders - Sends packets at line rate to benchmark packet processing performance of gateways and load balancers.
Packet Processing Speed Benchmarks - Measures packet processing speed by sending hundreds of millions of packets per second to evaluate raw throughput.
Stateless TCP Simulation - Creates and tears down TCP connections without full socket state, enabling billions of concurrent sessions.
User-Space Implementations - Implements a full TCP/IP protocol stack in user space to avoid kernel context switches.
Massive Connection Rate Generators - Creates and tears down TCP connections at massive scale to stress-test L4 load balancers and server capacity.
Traffic Generators - Generates tens of millions of HTTP connections per second and hundreds of Gbps throughput from a single x86 server.
Line-Rate Generators - Generates tens of millions of new connections per second and hundreds of gigabits of throughput to stress-test network infrastructure.
Multi-Protocol Generators - Produces mixed protocol traffic including HTTP, TCP, UDP, IPv4/IPv6, VLAN, and VXLAN for testing.
Real-Time Network Monitors - Prints per-second statistics including CPS, TPS, packet drops, socket errors, and HTTP status counts.
Connection Capacity Benchmarks - Benchmarks connection capacity by establishing billions of concurrent connections to measure maximum scale.
Concurrent Connection Simulation - Maintains an enormous number of simultaneous TCP or UDP sessions to evaluate connection-handling limits.
DPDK Line-Rate Generators - Uses DPDK to generate massive traffic volumes for stress-testing network devices and servers at line rate.
Multi-Protocol Traffic Orchestration - Creates mixed HTTP, TCP, UDP, IPv4/IPv6, VLAN, and VXLAN traffic to test network devices across diverse protocol scenarios.
HTTP Servers - Acts as a high-performance HTTP server or client to evaluate server and network performance.
Multi-Protocol Load Testing - Generates traffic across HTTP, TCP, UDP, IPv4/IPv6, VLAN, and VXLAN to test a wide range of network scenarios.
Multi-Architecture Compilation - Compiles DPDK drivers and application code for x86-64, ARM64, and LoongArch processors.
Traffic Pattern Simulation - Generates traffic to many destination IPs and supports symmetric RSS for accurate load distribution testing.
Per-Connection Latency Measurements - Records round-trip time for individual connections to assess responsiveness under load.
Traffic Distribution - Sends packets to a large set of target addresses to simulate realistic load distribution patterns.
Symmetric RSS Distribution - Generates traffic with hash-correct IPs for even load spreading across NIC RSS queues.
Multi-Architecture Support - Operates on X86-64, ARM64, and LoongArch processors to support diverse hardware environments.
Network Latency Testing - Measures round-trip time for generated traffic to evaluate target system responsiveness.
Per-Second Metric Aggregation - Aggregates and prints per-second metrics like CPS, TPS, packet drops, and HTTP status counts.
Per-Second Statistics Printers - Prints per-second statistics including connection rates, throughput, packet drops, and HTTP status counts.
Real-time Performance Monitoring - Prints per-second metrics like CPS, TPS, packet drops, and latency to observe system behavior under load.
HTTP Traffic Simulation - Simulates a high-performance HTTP server or client to evaluate server and network performance under load.
Network Performance Validation - Validates network performance across X86-64, ARM64, and LoongArch processors.
Standardized Performance Test Executions - Runs a suite of predefined benchmark tests including elephant flow, cloud PPS, and bandwidth measurement.

microsoft/ethr

5,861View on GitHub

ethr is a command line network diagnostic tool and performance analyzer designed to measure throughput, latency, and packet loss across TCP, UDP, and ICMP protocols. It functions as a multi-threaded network throughput tester and a diagnostic utility for analyzing connection health between client and server endpoints. The project differentiates itself through specialized diagnostics such as TCP retransmission monitoring and bidirectional path analysis to identify asymmetric performance differences. It provides the ability to evaluate the overhead of encrypted HTTPS connections and enforce spec

codesenberg/bombardier

6,797View on GitHub

Bombardier is a concurrent HTTP load generator and performance benchmarking tool written in Go. It functions as a latency distribution analyzer and benchmarking utility designed to measure the throughput and response speed of HTTP services by simulating high-volume request loads. The tool provides specific capabilities for TLS benchmarking, supporting client certificate authentication and the ability to bypass server certificate verification. It distinguishes itself through the use of a token-bucket algorithm for precise request rate limiting and the use of templates to export benchmark resul

firehol/netdata

79,416View on GitHub

Netdata is a real-time infrastructure monitoring tool and multi-node observability platform. It functions as a high-resolution monitoring agent, log and metric aggregator, and time-series database designed to provide full-stack visibility into server health. The system is distinguished by its per-second metric sampling and zero-configuration auto-discovery, which allows for immediate infrastructure tracking upon installation. It utilizes edge-based machine learning and unsupervised models to detect system anomalies and abnormal metric patterns locally on each node. For distributed environment

libnyanpasu/clash-nyanpasu

13,040View on GitHub

Clash Nyanpasu is a network proxy client and graphical user interface for managing Clash network cores. It functions as a proxy manager and traffic monitoring dashboard, allowing users to route system traffic and handle proxying through a visual interface. The project distinguishes itself by providing a TUN mode proxy manager for controlling system-wide network tunneling. It features a core-agnostic API layer that standardizes communication between the interface and various interchangeable network core implementations. The application covers proxy configuration management, including profile

How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.

baidudperf

Dperf

Features