AUTODAPS: Automatic Topology Analysis for Distributed Anomalies Prevention Systems in the IoT
1. AUTODAPS: Automatic Topology Analysis for Distributed
Anomalies Prevention Systems in the Internet of Things
Internet of Things
Abstract
Blockchain
Botnets&DDoS
Internet of Things and Blockchain will soon become the building blocks of future technologies. The
immense power of this combination is very attractive for those who want to perform malicious activities.
As such, is of paramount importance to tackle this threat in a timely manner. AUTODAPS aims at
investigating the security of this new paradigm with the most advanced techniques available to date.
Infected devices connected to the
Internet can form a Botnet controlled
by a malicious user. Botnets are used
to perform massive malicious
activities such as Distributed Denial of
Serivice (DDoS) or spamming.
The Internet of Things (IoT) is
the network composed by all
smart devices connected to
the Internet.
They are able to communicate
with each other and they each
have their own computing
power.
A Blockchain is a shared
database (DB) controlled by
nodes of a peer-to-peer network.
Each node has a copy of the DB,
can add new records (or
transactions), and it eventually
agree with others on the validity
of the data.
● Develop a IoT protocol on top of the
Bitcoin blockchain
● Simulate Botnet activities
● Apply TDA+ML techniques to study
topology changes due to Botnets
Future Work
Context
AUTODAPS project is in
collaboration with Nokia
Bell-Labs Paris and co-fundend
by the Maria de Maeztu
Excellence program.
AUTODAPS Homepage:
Vanesa Daza
vanesa.daza@upf.edu
Matteo Signorini
matteo.signorini@nokia-bell-labs.com
IoTBotnets
According to recent statistics, there
will be 50 billion devices by 2020.
This could give an attacker a huge
power for carrying out malicious
activities.
Approach
TDA&MLQuestion
We developed a tool, for the Bitcoin-based
Colored Coins protocol, that retrieves all
transactions involving a specific asset and
builds a graph, where nodes represent asset
holders and edges represent transactions.
TransactionGraphCommunicationover
Bitcoin
Topological Data Analysis (TDA)
techniques allow to visually extract
information about big amounts of data
(Big Data) from the topology of a graph
representing such data. By analyzing
the graph of communications in an IoT
network, and leveraging Machine
Learning (ML) techniques, it is possible
to create new algorithms to detect
botnet activities and take the
appropriate countermeasures.
The huge amount of devices
connected to the Internet pose a
serious threat to our security
We want to use Machine Learning
and Topology Data Analysis
techniques to detect botnets in
blockchain-based IoT
We developed a framework for
exchanging messages over the Bitcoin
network. Messages are split into chunks
and embedded into transactions.
Future IoT networks will be based on blockchain technologies
We developed the building blocks for
the simulation and analysis of a
blockchain-based IoT network
Federico Franzoni
federico.franzoni@upf.edu
Acknowledgments
Motivation Preliminary Results
Conclusions