Home

Awesome

NuGet Build status

New!

The newly released PacketDotNet.Connections is a supplement to Packet.NET that adds connection tracking and http following support.

Packet.Net

Packet.Net is a high performance .Net assembly for dissecting and constructing network packets such as ethernet, ip, tcp, udp etc.

Originally created by Chris Morgan chmorgan@gmail.com

https://github.com/chmorgan/packetnet

Code is found in the PacketDotNet namespace.

Performance

Packet.Net has been designed for the highest performance possible. As such we aim to perform the most minimal amount of data processing in order to fully determine the datagram nesting.

For example a TCP packet would be parsed into a series of linked objects like: Ethernet -> IPv4 -> TCP but no further data processing is performed until particular fields are accessed. In addition the objects point to packet memory in-place, avoiding allocation and copying of the packet contents unless necessary, such as when altering data payloads or resizing variable length fields.

Test suite

Packet.Net has a comprehensive suite of tests for each of the supported packet types, see the 'Test' subdirectory.

Supported packet formats

Capture example

See Capturing packets example

<p align="center"><img src="/terminalizer/captureexample.gif?raw=true"/></p>

Getting started

A few basic examples can be found in the Examples/ directory.

Debug vs. Release builds

The Debug build depends on log4net and has log4net calls in some of its classes and code paths.

The Release build does NOT depend on log4net and, taking advantage of conditional method attributes, does not include any calls to log4net methods. This ensures that there is no performance impact on release builds.

Performance benchmarks

The Test/ directory contains a few benchmarks that were used to guide the design and implementation of Packet.Net. These benchmarks either contain 'performance' or 'benchmark' in their names.

If you have a performance concern or issue you'll want to write a concise test that reproduces your usage case in a controlled manner. It will then be possible to run and re-run this test case in various profiling modes in order to look at potential ways of optimizing code. The tests will also provide a baseline from which to compare any proposed performance improvements in order to ensure that changes are not inadvertantly reducing instead of increasing performance.