Network Traffic Flow Detaction
Traffic Identification and classification techniques such as DPI (Deep Packet Inspection) are used widely, which prove to be expensive in terms of latency and other viable resources. We are working on a technique using much lower resources based on Machine Learning. As an example, the Global Interne
2025-06-28 16:28:39 - Adil Khan
Network Traffic Flow Detaction
Project Area of Specialization Artificial IntelligenceProject SummaryTraffic Identification and classification techniques such as DPI (Deep Packet Inspection) are used widely, which prove to be expensive in terms of latency and other viable resources. We are working on a technique using much lower resources based on Machine Learning. As an example, the Global Internet Traffic Statistics tell us that 15% of all Internet Traffic is acquired by Netflix alone, followed by YouTube, which is responsible for 11%. Streaming being a big chunk of the traffic takes the majority fill of the internet pipe of major Internet Service providers. Many functional tools and methods have been introduced in the past to detect an upcoming flow so the traffic can be throttled using policing and shaping, but the detection techniques are mainly based on DNS and port numbers, which can easily misidentify new traffic sources and is against net neutrality principles as well.
Project ObjectivesFlow Detection and classification on a packet Dump ( Pcap ) based on signatures irrespective of the origin of the source.
Project Implementation MethodCollection Of Data:
- Collected network traffic data using wireshark tool of computer.
- Collected network traffic data of android phone using PCAPdroid application
Data Cleaning through EDA:
Performed multiples steps to clean data in our data set using EDA cleaning methedologies.
Algorithms:
Implemented multiples algorithms on our final data set to get the best accuracy by comparing the results.
Benefits of the ProjectUse Cases:
- we can control the bandwidth of the network pipe.
- we can do debugging where the network is getting disturbed.
Flow Detection and classification on a packet Dump ( Pcap ) based on signatures irrespective of the origin of the source.
Final Deliverable of the Project Software SystemCore Industry ITOther Industries Telecommunication Core Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development GoalsRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 59000 | |||
| Nividia GT 730 4gb | Equipment | 1 | 24000 | 24000 |
| core i5 7th generation | Equipment | 1 | 35000 | 35000 |