Stream Management on a Cloud

Description: This research focuses on stream management on a cloud that enables the storage, retrieval, searching and processing of streaming data. The projects briefly outlined below comprise this collaborative research with Telus:

  • Indexing Techniques for “Content Addressable” Streams: Data streams, for example those generated from twitter feeds or IoT devices that give rise to sensor data often need to be stored for a detailed analysis at a later point in time. Effective indexing techniques that will lead to fast searching of this data are under investigation.
  • Techniques for Detecting Complex Events: Multiple events occurring within a short period of time on separate independent streams coming from multiple different IoT devices for example, often need to be processed simultaneously to detect the occurrence of a “complex” event. This project focuses on a method and system for complex event processing for remote patient monitoring.
  • Resource Management Techniques for Enabling Streaming Data Analysis: Frameworks such as Apache Storm are available for performing the analysis of streaming data. Associating priorities with data streams and devising priority based resource management techniques for Storm form the focus of attention for this project.

Sponsor: Telus, NSERC
Date: 2016-2020
Members: Shikharesh Majumdar, Marc St-Hilaire, Ali El-Haraki (Telus)

Machine Learning Techniques for Reliable Multicast Routing in Mobile Ad-Hoc Networks

Description: Multicast routing refers to the transmission of packets to a group of nodes identified by a single multicast group address. It plays a critical role in supporting applications that require group communication such as file distribution. One particularly challenging environment for multicast routing is Mobile Ad-hoc Networks. The major problems facing routing in such networks are node mobility, frequently changing topology, unstable wireless links, and limited transmission range.

Recently, Machine Learning (ML) has emerged as a major field of research for routing in wireless ad-hoc networks. ML is an application of artificial intelligence that enables a system to learn, act, improve its learning over time from interacting with the environment, and automatically adapt to any changes in the environment. As a result, ML techniques can be used to dynamically adapt routing paths to cope with any changes such as link failure, interference and congestion. In this project, we aim to evaluate the performance of applying machine learning strategies to multicast routing in Mobile Ad-hoc Networks and compare it with the performance of existing multicast routing protocols.

Sponsor: CRC
Members: Thomas Kunz and Marc St-Hilaire


Description: Modern building automation systems collect and store vast amounts of data in temporal scales ranging from seconds to multiple years and spatial scales ranging from rooms to communities. But tools that facilitate management and visualization of these data are in their infancy. It has become abundantly clear that data availability alone has not significantly improved the way buildings operate. Current visualization methods for buildings are a bar or line graphs, car dashboard displays (e.g., dials), and pie charts. These methods fail to communicate the complexities of building performance and equal opportunities to improve performance.

We are defining new methods and tools for integrating real-time sensor data with occupancy information based on the occupants’ devices (smartphones, tablets, laptops). And, integrate parametric data from the building to construct formal models that can be used for both simulation and control. The formal models will provide analytic results that will be then deployed in advanced 3D visualization environments (including a full campus map model). We will conduct real experimentation in the four test buildings at Carleton University, and at MaRS (Autodesk Research offices). The results will be used by Autodesk Research in their BIM tools to allow Architecture, Engineering, Construction, and Operations(AECO) professionals to analyze real-time data and visualize the modeling results and the analysis results into their actual designs, improving the overall process.

Sponsor: NSERC, Autodesk
Members: Gabriel Wainer

Mathematical Techniques for Enhancing Security and Privacy

Description: Apply mathematical techniques from fields such as Game Theory and Optimization to improve the levels of security or privacy in organizations, computer systems, or critical infrastructure. Game theory can provide insight into attacker behaviour that can help in predicting attack patterns and formulating an effective defence. Optimization techniques can be used to guide the allocation of scarce resources to improve the security level.

Sponsor: Aptusinnova Inc.
Date: 2019-2022
Members: George Yee

Advanced Methods in Information Security and Privacy

Description: Investigate new techniques for assessing and enhancing information security and privacy protection. New ideas investigated to date include the development of metrics for assessing security level, and optimizing the application of security measures to vulnerabilities in order to achieve a target security level based on the metrics. Future research in this area includes the assessment of security level for critical infrastructure and new paradigms for sharing health information that preserves privacy.

Sponsor: Aptusinnova Inc.
Date: 2019-2022
Members: George Yee

Building Software that Safeguards Security and/or Privacy

Description: Investigate methods and tools for building software that is secure and/or respects and protects privacy. The goal is to build-in security and privacy right from the beginning of software development. As a consequence of this goal, the methods and tools addressed to date apply to the architectural design level. The research has focused on the use of models to identify risks and vulnerabilities in a software system. Future research in this area includes more effective methods for threat modelling.

Sponsor: Aptusinnova Inc.
Date: 2019-2022
Members: George Yee