Cistel has taken the challenge of embracing and conducting scientific research in various cutting-edge networking areas and has been working closely with best-in-class Canadian researchers from Dalhousie University, Carleton University, University of Ottawa, and Concordia University, to advance new technologies in these areas.
As a technology and management services company, Cistel not only provides managed IT services for government, public and private organizations but also manages IT infrastructure for several medium and large organizations in the cloud computing environment. Thus, the ability to provide consulting and support on information security management in the cloud computing environment is one of the main goals of Cistel. To this end, Cistel is involved in the research and development of a cloud-hosted system for network security monitoring and anomaly detection. The R&D project aims to investigate the best practices for security monitoring systems based on machine learning and deep learning anomaly detection for such environments. Also, Cistel is working in collaboration with researchers from Dalhousie University and the University of Waterloo on R&D projects aiming at developing end-to-end security models for Supervisory Control and Data Acquisition (SCADA) systems and IoT based monitoring and control systems.
Natural Language Processing
In this research project, the aim is to develop multiple NLP-based machine learning models that would allow information retrieval from given documents and databases. Word embedding, such as Word2Vec, GloVe, ELMo, and BERT based models, are being evaluated with fine-tuning for a smart office assistant which is a query-response application for the documents. For database query applications, question processing and SQL query formation for a pre-defined database with known schema will be performed using the NLP technique. A prototype smart office assistant will be eventually developed that will demonstrate the effectiveness of machine learning techniques, more specifically the effectiveness of natural language models for a smart office application for information retrieval from corporate document repository (e.g. Share Point) and database. This is a collaborative R&D project between Carleton University and Cistel Technology.
Voice and Multimedia Services over IP
Cistel has conducted research into methodologies and knowledge-based tools which can be used to develop, test and manage Voice Over IP (VoIP) gateway products. The vision and design of a knowledge-based software tool mainly for testing VoIP gateway applications is in progress. The rapid technological advancement in many telecommunication areas such as in the area of voice over IP creates an enormous challenge to the tester. Effective and efficient testing depends on the knowledge of the testers on the technology to a great extent and it is extremely challenging to keep up with the new technology with the time to market pressure and tight delivery schedule. In a unique way, the tool acts as a subject matter expert in VoIP and as an experienced tester helping junior testers and/or testers new to this field develop excellent test plans with suites of test cases to test various functionalities of the VoIP gateway products.
Cistel has also been looking into methodologies to test voice/video quality in a simulation test bed. This would enable research in voice and video over IP quality. Extensive research has been also done towards providing cost-effective VoIP PBX solutions to Small to Medium size Businesses (SMB).
Bridge Infrastructure Management
Bridge Infrastructure Management using the advanced Information and Communication Technology (ICT) infrastructure is extremely important. Currently, the bridge infrastructure problem is addressed in a compartmentalized fashion. Solutions are researched in isolated areas of design, maintenance and repair, monitoring, asset management, policy, and planning. But, the existing approach in bridge engineering practice neither facilitates the transfer of information and knowledge, nor does it encourage collaboration among bridge engineers and researchers, owners, or other stakeholders. Effective monitoring, management, and maintenance of these infrastructures are important for the safety of the public as well as for reducing the cost associated with the operation and maintenance.
In recent years, the progress in information and communication technology, which includes cost-effective ubiquitous network technology and the availability of various distributed system engineering techniques (e.g. Service Oriented Architecture (SOA) tools), provides tremendous opportunity to address these complex problems in a distributed and cost effective manner. Essentially, it is necessary to network, enable, and manage various resources such as sensors, computing and storage devices, databases of historic data regarding the state and maintenance of the infrastructure, and software tools so that these resources can be remotely accessed and shared through various user interfaces. The key to the solution of the infrastructure management problem is middleware that will provide the connectivity and inter-operability among diverse resources and manage these resources in an effective manner. Although the focus of the research is on a middleware that will integrate a variety of different geographically dispersed resources for improving the current practice in bridge management in a cost-effective manner, the middleware is envisioned to be generic in nature and can be adapted to other infrastructure management problems with a modest effort.
Wireless Sensor Ad-hoc Networks
This research specifically aims at developing a suite of novel algorithms which can be used in sensor wireless networks for routing information about the nodes which are critical for wireless communications. On a broader scale, the research approach is also towards development of an advanced position based routing, allowing a user to make peer-to-peer multimedia communications without any infrastructure available. An on-demand network is to be established that is secure, robust and reliable in realistic physical wireless media which is prone to limited channel capacity, dynamic and changing positions of nodes, signal degradation and fading, nodes going to sleep mode to preserve power and interference of the signal. The sophisticated position based networking layer routing will allow the users of existing MAC layer and link layer nodes to establish and reliably forward multimedia traffic in this hostile and dynamic environment. The advanced features of the routing algorithm will be such that the various requirements are properly met including being:
- Reliable – there should be always reachability, the algorithm must be guaranteed to find a route, if any exist;
- Fault-Tolerant – nodes in an ad-hoc network can come and go at any point in time or malfunction, but the routing algorithm must work;
- Energy Efficient – nodes cannot be computationally too intensive; need to ensure low energy consumption; and
- Secured – the algorithm must take into account the security aspect as an ad-hoc network is more vulnerable to attacks.
Developing and characterizing a suite of novel algorithms would definitely help in realizing the ad-hoc networks with sensors for different practical applications. These algorithms will be developed such that they can be easily applied in various different applications given the requirements. The scientific and technology advancements in this research would lead to many practical applications where an optimum routing solution is a must. Different case studies are also conducted to validate these algorithms against different practical applications.