NetPulse

Lightweight Network Bandwidth & Latency Analyzer

📌 Projects Included

1. Latency & Bandwidth Analyzer

Analyzes latency distribution (avg, min, max, p95, p99) and bandwidth.

2. TCP Latency Checker

Measures round-trip time (RTT) at microsecond precision.

---

🧠 Key Focus Areas

• Low-level networking (POSIX sockets)
• Latency measurement & analysis
• Tail latency (p95/p99)
• System-level performance behavior

---

📁 Folder Structure

└── 📁NetPulse
    └── 📁Latency_and_Bandwidth_Analyzer
        └── 📁bin
            ├── client
            ├── server
        └── 📁client
            ├── client.cpp
        └── 📁server
            ├── server.cpp
        └── 📁utils
            ├── protocol.h
            ├── timer.h
    └── 📁TCP_latency_checker
        └── 📁bin
            ├── client
            ├── server
        └── 📁client
            ├── client.cpp
        └── 📁server
            ├── server.cpp
    ├── .gitignore
    ├── Notes.txt
    └── README.md

1. Network Latency & Bandwidth Analyzer

A C++-based network performance analysis tool that measures latency and bandwidth using POSIX sockets.

---

🚀 Overview

This project extends basic latency measurement by analyzing:

• Average latency
• Minimum & maximum latency
• Tail latency (p95, p99)
• Bandwidth (Mbps)

It provides a more realistic view of network performance by focusing on latency distribution, not just averages.

---

⚙️ Tech Stack

• C++
• POSIX Socket APIs
• TCP/IP Networking
• High-resolution timers

---

🧠 Key Concepts

• Latency distribution (avg, min, max)
• Tail latency (p95, p99)
• Bandwidth estimation
• Performance variability (jitter)
• Throughput vs latency trade-offs

---

🛠️ Build & Run

1. Compile

mkdir -p bin
g++ client/client.cpp -Iutils -o bin/client
g++ server/server.cpp -Iutils -o bin/server

2. Run

Terminal 1 (Server):

./bin/server

Terminal 2 (Client):

./bin/client

---

📊 Sample Output

--- Network Analyzer Report ---

Latency (ms):
  avg: 3.12702
  min: 0.235
  max: 5.604
  p95: 5.537
  p99: 5.597

Bandwidth:
  8.36027 Mbps

📸 Sample Output

🔍 Observations

• Average latency does not reflect real performance

• Tail latency (p95/p99) is significantly higher than average

• Latency spikes indicate:

  • OS scheduling delays
  • Buffering and TCP overhead

• Bandwidth depends on data transfer efficiency and system limits

---

🎯 Learning Outcome

• Importance of latency distribution in real systems
• Understanding of tail latency (critical in HFT systems)
• Practical experience in network performance measurement
• Insight into TCP behavior and overhead

---

🚀 Future Improvements

• UDP-based analyzer for comparison
• Non-blocking I/O for higher performance
• Parallel/multi-client testing
• Visualization dashboard (graphs)

---

2. TCP Latency Checker

A lightweight C++ client-server application built using POSIX sockets to measure round-trip time (RTT) at microsecond precision.

---

🚀 Overview

This project measures network latency by sending small messages between a client and server over TCP and calculating the time taken for each round trip.

It is designed to demonstrate low-level networking, latency measurement, and system-level variability.

---

⚙️ Tech Stack

• C++
• POSIX Socket APIs (socket, connect, send, recv)
• TCP/IP Networking
• High-resolution timing (std::chrono)

---

🧠 Key Concepts

• Round Trip Time (RTT)
• Latency measurement in microseconds
• TCP communication overhead
• OS scheduling impact on latency (jitter)

---

🛠️ Build & Run

1. Compile

mkdir -p bin
g++ client/client.cpp -o bin/client
g++ server/server.cpp -o bin/server

2. Run (in two terminals)

Terminal 1 (Server):

./bin/server

Terminal 2 (Client):

./bin/client

---

📊 Sample Output

==== Latency Stats ====
Packets Sent: 100
Avg RTT: 468 microseconds
Max RTT: 2945 microseconds

📸 Sample Output

---

🔍 Observations

• Latency is very low on localhost due to absence of physical network
• However, latency is not constant and shows jitter
• Variability is caused by:
  • OS scheduling
  • Context switching
  • Background processes

---

🎯 Learning Outcome

• Hands-on experience with POSIX socket programming
• Understanding of latency behavior and variability
• Insight into system-level effects on performance

---

🚀 Future Improvements

• Add UDP comparison (TCP vs UDP latency)
• Introduce non-blocking sockets (select() / poll())
• Multi-client support
• Real network testing (different machines)

---