Course Description

• Teach common network programming techniques by examples
• Focus on server-side TCP network programming on Linux
• Walk-through code in C++, Python and Go

Prerequisites

• Have read Unix Network Programming volume 1 (either 2nd ed. or 3rd ed.) by late W. Richard Stevens
• Know basics of TCP/IP, eg. IP addresses, ports,
• Know how to write an echo server with basic sockets APIs in any language

Code

• http://github.com/chenshuo/muduo
• http://github.com/chenshuo/recipes
• http://github.com/chenshuo/muduo-protorpc
• http://github.com/chenshuo/muduo-examples-in-go

Schedule

0. Overview

• Measure throughput of TCP over Gigabit Ethernet with netcat
• Lecture section 1 | Lecture section 2 | Lecture section 3

Part I: Basic non-concurrent programs with blocking IO

1. TTCP

• Review basic Sockets API
• Measure throughput of TCP over Gigabit Ethernet with TTCP
• TCP self connection
• Lecture section 4 | section 5 | section 6 | section 7 | section 8 | section 9

2. Round-trip

• How NTP works
• Measure clock error between two hosts
• The only UDP example
• Lecture section 10 | section 11 | section 12 | section 13 | section 14 | section 15 | section 16

3. Netcat

• The Swiss army knife of networking
• Deal with two file descriptors simultaneously (stdin/out and TCP socket)
• Why IO-multiplexing must be used with non-blocking IO
• Lecture section 17 | section 18 | section 19 | section 20 | section 21 | section 22 | section 23

Part II: Concurrent programs with non-blocking IO

4. procmon

• cpu / memory usage charts of one process
• dummyload
• Lecture section 25 | section 25 | section 26 | section 27 | section 28 | section 29

5. Simple memcached

• Customize the data structure
• Efficient locking by sharding
• Performance testing
• Lecture section 30 | section 31 | section 32 | section 33 | section 34 | section 35 | section 36 | section 37 |

6. Sudoku Solver

• Request-response model
• Client connections are independent
• Concurrency and parallel on multicore machines
• Disable Nagle's algorithm in low-latency servers
• Load balancing
• Overload protection
• Lecture section 38 | section 39 | section 40 | section 41 | section 42 | section 43 | section 44 | section 45 | section 46 | section 47 | section 48 | section 49 | section 50 |

7. Broadcasting to TCP peers

• Connections exchange data with each other
• How to deal with slow clients?
• Lecture section 51 | section 52 | section 53 | section 54 | section 55 | section 56 |

8. SOCKS proxy server

• Relay two TCP connections
• Mismatching bandwidth?
• Difference between level 4 and level 7?
• Lecture section 57 | section 58 | section 59 | section 60 | section 61 | section 62 | section 63 | section 64 | section 65 | section 66 | section 67 | section 68 |

Part III: Data processing with multiple machines

9. Parallel N-queen puzzle

• Multithreaded N-queen puzzle
• Multi-machine N-queen puzzle using MapReduce
• Lecture section 69 | section 70 | section 71 | section 72 | section 73 | section 74 |

10. Median across machines

• Lecture section 75 | section 76 | section 77 | section 78 |

11. Most frequent queries

• Lecture section 79 | section 80 | section 81 | section 82 | section 83 | section 84 |