Research database

Deep-Augur - INF (Digital Infrastructure)

Duration:
12 months (2018)
Principal investigator(s):
Project type:
UE-funded research - EIT
Funding body:
COMMISSIONE EUROPEA
Project identification number:
18217

Abstract

The traditional Deep Packet Inspection (DPI) based traffic analysis solutions rely on accessing packet payload to identify network flows and require massive amounts of resources (HW and SW). The Deep Augur Smart Traffic Analysers facilitate and speed up the characterisation of network flows by means of advanced Deep Neural Network techniques. The current traffic analysis products rely on highly resource consuming inspection of packet payload. The Deep Augur smart traffic analysers (STA) are based on deep learning techniques and can analyse network flows (classify, predict, detect anomalies, etc.) without accessing the packet payload. The Deep Augur generator trains generic machine learning components and produces smart traffic analysers tailored to solve the problems of the customer network.

Structures

Partners

  • UNIVERSIDAD POLITECNICA DE MADRID

Budget

Total cost: € 553,414.00
Total contribution: € 415,059.00
PoliTo total cost: € 155,000.00
PoliTo contribution: € 116,250.00