# PacketBreeze
PacketBreeze is an utility tool built for analysing and classifing packets by looking at packet headers. 

## Primary design goals:

PacketBreeze aims to:

 - **Classify packets and flows as benign or malicious with high true positives (TP) and low false positives (FP)**. 
 - **Use the labeled data to reduce amount of traffic requiring deeper analysis**. 

Additionally, PacketBreeze also categorizes packets into flows and shows a rich ensemble of flow data and statistics.

| <img width="1559" alt="Packetbreeze-flows" src="assets/images/packetbreeze.png"> |
|:--:| 
| *PacketBreeze takes packets and returns file with statistics of flows.* |



| <img width="400" alt="Packetbreeze-Confusion-Matrix" src="assets/images/confusion.png">
|:--:|
| *Packetbreeze's output shown in a confusion matrix.*  |

## When to use Packetbreeze

Use PacketBreeze if you wish to build and operate machine-learning models on network packet data.

## Quick Start

Start by cloning the project: [link](https://git.weirdnatto.in/Unic-X/PacketBreeze.git)

Build the binary:



```console
foo@bar:~$ go build -o packetbreeze
```

Run the Binary:

```console
foo@bar:~$ ./packetbreeze -ifLiveCapture=true -fname=webgoat -maxNumPackets=40000000 -ifLocalIPKnown false 
```

<p>
	<em>Offline analysis of the PCAP packets</em>
</p>

---

### Requirements

Go
Python Requirements in [File](assets/requirements.txt)



## Who uses PacketBreeze?

 * One can use PacketBreeze to label the network packets using ML before deep analysis. Thus resulting in faster analysis overall.

 * Overall the target users for this project are intermediate network analyst, and network engineers.

## Get in touch

Thank you for using PacketBreeze

Community contributions are always welcome through GitHub Issues and Pull Requests.



## Developers

 Arman Singh Kshatri `221020412`
 Swastika Satya	`221020453`