ADON 2018 : International Workshop on Anomaly Detection ON the Cloud and the Internet of Things
Call For Papers
ADON 2018: International Workshop on Anomaly Detection ON the Cloud and the Internet of Things,
December 10, 2018, as part of
IEEE CloudCom 2018, 11-13 December 2018, Hilton Cyprus, Nicosia, Cyprus
Anomalies are detected in systems as a result of malicious behavior of users or as unscheduled changes in the operation of a system. With the advent of cloud, similar behavior is now detected in virtualized environments such as the environment of a cloud provider (now affecting the operation of the system in scale and of a much large number of users) with certain economic and operational impact. Although cloud systems are considered to be more efficient, for example in terms of reliability, security etc. compared to legacy systems operating within the premises of a company, they are exposed to a much larger number of users and the internet. At the same time, due to its scalability and affordability, the cloud is considered to be the ideal environment for deploying IoT applications. This exposes the cloud to even more risks as IoT is operating in the periphery of the cloud and is generally less protected than the cloud itself. In particular, the advent of the cloud and Internet of Things (IoT) open-up new possibilities in the design and development of methodologies ensuring reliable security protection and, in the case this fails, of methodologies for detecting and for dealing with the cause and point of system failure.
Due to the size and complexity of modern systems, anomalies can be detected in many aspects of system operation and relate mainly to:
* Anomaly detection for malicious behavior detection which is typically expressed as (a) Fraud detection in which case, authorized of unauthorized users operate the system for the purpose of unfair or unlawful gain and (b) Intrusion detection in which case, unauthorized users are attempting to disrupt normal system operation.
* Anomaly detection on large scale system failures which is due to heavy (CPU, network and memory) workloads or faulty/misconfigured resources. A special case of system failure is encountered when parts of the system fails to operate as scheduled due to power failure or material fatigue (e.g. disk failure).
* Anomaly detection on IoT systems is due to unexpected behavior of connected devices which can be detected by monitoring the operation of these devices on the network, or by the analysis of real time data streams of misconfigured devices, or by monitoring and analyzing network traffic.
Anomaly detection has been studied extensively in recent years and new methods are now becoming available on the cloud. Depending on application, anomalies can be detected either in real time i.e. typically by the analysis of stream data acquired by the application and operation of the system or, in batch (i.e. by analyzing system log data). Methods and systems for stream processing for example Storm, Spark, Flink, big data analysis techniques (as log data eventually become big) combined with Machine Learning techniques (for adapting anomaly detection to the peculiarities of the data and of the operation environment) are of particular importance to the design of anomaly detection methods. Combined with methods of system security analysis in virtualized environments (such as the cloud), the new era of methods for anomaly detection will soon arise.
The purpose of this Workshop in to bring together experts from the fields of distributed computing systems including security, cloud and Internet of Things as well as experts on algorithms for signal processing, log analysis, pattern recognition and statistical learning models, working in all aspects of anomaly detection such as those referred to above.
Topics of interest:
Methods and Tool for anomaly detection in virtualized (including containerized) environments and the IoT with particular emphasis to:
* Malicious behavior detection (fraud detection, intrusion detection)
* Security breach detection in IoT
* System failure detection
* Stream processing for anomaly detection (real time and batch processing techniques)
* Stream query processing for anomaly detection
* Service oriented architectures for anomaly detection
* Machine learning platforms and techniques for anomaly detection
* Use cases and performances of anomaly detection methods
* Application of machine learning methods for anomaly detection
* Pattern-based and time-series system data analytics for anomaly detection
* Log data analysis and data mining for anomaly detection
* Text mining and regular expressions processing for anomaly detection
* Black-box anomaly detection techniques for detection of erroneous events
Paper submission: 31 August 2018
Notification of acceptance: 14 September 2018
Camera-ready version: 08 October 2018 (please observe main CloudCom 2018 website)
Workshop day: December 10, 2018.
Submitted papers must not substantially overlap papers that have been published or that are simultaneously submitted to a journal or a conference with proceedings. Authors must submit their papers by the deadline indicated above, using the Easychair submission system. Only PDF files will be accepted. Manuscripts need to be prepared according to the IEEE CS format. Only PDF files will be accepted. All regular paper submissions should be written in English with a maximum paper length of 6 pages, two-column (same as the main conference). All submitted papers will be reviewed by at least three experts.
Accepted papers must be presented at the conference. At least one author of each accepted paper must register to the conference, by the early date indicated by the organizers, and present the paper.
Accepts papers will be included in the conference proceedings of CloudCom 2018, will be published by IEEE Conference Publishing Service with ISBN and ISSN, and submitted to IEEE Xplore and indexing services such as Ei Compendex.
The registration policy of the main conference applies to the workshop. At least one author of each accepted paper must register to the main conference. There will be 1 fee for the whole conference that covers 1 paper. The authors should be referred to the conference registration page when this opens.
Euripides G.M. Petrakis, Technical University of Crete, Greece
Stelios Sotiriadis, Birkbeck, University of London, UK
Program Committee (in alphabetical order)
Cristiana Amza, University of Toronto, Canada
Nik Bessis, Edge Hill University, UK
Rajkumar Buyya, University of Melbourne, Australia
Antonios Deligiannakis, Technical University of Crete, Greece
Francesco Iorio, Autodesk, Canada
Florin Pop, Technical University of Bucharest, Romania
Alessandro Provetti, Birkbeck, University of London, UK
Sahil Suneja, IBM T.J. Watson, Toronto, Canada
Marcelo Trovati, Edge Hill, UK