A5047687728- Software-Defined Networks and 5G
- IoT and Edge/Fog Computing
- Energy Efficient Wireless Sensor Networks
- Mobile Ad Hoc Networks
- Opportunistic and Delay-Tolerant Networks
- Cloud Computing and Resource Management
- Indoor and Outdoor Localization Technologies
- Network Security and Intrusion Detection
- Advanced Optical Network Technologies
- Underwater Vehicles and Communication Systems
- Internet Traffic Analysis and Secure E-voting
- Network Traffic and Congestion Control
- Caching and Content Delivery
- Advanced Malware Detection Techniques
- Energy Harvesting in Wireless Networks
- Advanced Data Storage Technologies
- Data Stream Mining Techniques
- Wireless Networks and Protocols
- Cooperative Communication and Network Coding
- Smart Grid Security and Resilience
- Bluetooth and Wireless Communication Technologies
- Image and Video Quality Assessment
- Video Coding and Compression Technologies
- Smart Grid Energy Management
- Speech and Audio Processing
Dalhousie University
2017-2025
University of California, Riverside
2016
University of Alberta
2006-2015
Concordia University
2005
Nokia (United States)
1991
One of the primary architectural principles behind Internet is use distributed protocols, which facilitates fault tolerance and management. Unfortunately, having nodes (i.e., switches routers) perform control decisions independently makes it difficult to network or even understand debug its overall emergent behavior. As a result, networks are often inefficient, unstable, fragile. This architecture also poses significant, insurmountable, challenge deployment new protocols evolution existing...
In this paper, we consider the indoor localization problem, i.e., identifying Cartesian coordinates of an object or a person under roof. To solve problem RF-based method called profiling, two-step process, where radio map monitored area is first constructed by collecting signal strength from known locations. An unknown location then predicted using as reference. propose K nearest neighbor (KNN) profiling-based dubbed LEMON (location estimation mining oversampled neighborhoods). It based on...
Bias in machine learning has rightly received significant attention over the past decade. However, most fair (fair-ML) works to address bias decision-making systems focused solely on offline setting. Despite wide prevalence of online real world, work identifying and correcting setting is severely lacking. The unique challenges environment make addressing more difficult than First, Streaming Machine Learning (SML) algorithms must deal with constantly evolving real-time data stream. Secondly,...
Software-defined networking (SDN) is a new approach to designing networks. SDN decouples and migrates network control from the hardware, which enables innovative efficient design operation. The SDN-based well adopted in data centers enterprises. In this paper, we state that beneficial multihop wireless networks like sensor (WSN) propose an architecture called SDSense. We novel principled of WSN design, where decompose functions as slow (e.g., topology control) fast congestion changing...
The increasing use of Machine Learning (ML) software can lead to unfair and unethical decisions, thus fairness bugs in are becoming a growing concern. Addressing these often involves sacrificing ML performance, such as accuracy. To address this issue, we present novel counterfactual approach that uses thinking tackle the root causes bias software. In addition, our combines models optimized for both performance fairness, resulting an optimal solution aspects. We conducted thorough evaluation...
Programmable accelerators enable the execution of applications intended for running in usual servers. However, inappropriately on these devices can lead to load imbalance and performance degradation. An alternative tackle this problem is balancing, but existing in-network balancers typically have no visibility often hard code policies switch source code. In article, we present <sc xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">P4Hauler</small> , an...
Pivoting remains an economical and practical penetration method as it allows a malevolent actor to obtain access private network through compromised devices. There are various tools both on the web native many operating systems, making pivoting simple execute, even with limited system access. Preventing these attacks is traditionally performed detection software running end hosts or perimeter devices, e.g., firewalls. However, not all end-host devices under administrator control, attackers...
Software-Defined Networking (SDN) can help simplify the management of today's complex networks and data centers. SDN provides a comprehensive view network, offering flexibility easing automation. In SDN, traffic functionality requires high-performance responsive controller. this paper, we conduct an experimental evaluation two open-source distributed OpenFlow controllers, namely ONOS OpenDaylight. Specifically, construct testbed use standard benchmarking tool called Cbench to evaluate their...
The "first wave" of wireless ad hoc routing protocols was heavily influenced by the algorithms used in wired networks. However, their overheads due to topology changes and implicit restriction point-to-point control messages meant that extensive effort had be expended "fixing" them fit intricacies opportunities offered environments. In this paper, we analyze compare most widely mobile networks (MANETs). particular, consider overheads, storage requirements, network setup costs these see how...
An overview of different self-healing network architectures, focusing on ring is presented. Digital cross-connect systems-based and diversely routed protection are briefly discussed. Since a typical design may use all the discussion applications perspectives, illustrating how various technologies best fit in evolving network, included.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>
In software-defined WANs (SD-WAN), link failure can lead to congestion and packet loss, hence degrading application performance. State-of-the-art traffic engineering approaches speed up recovery by proactively installing backup tunnels redirecting affected immediately in the data plane, which reduces burden on network controller. However, these either bandwidth waste because of reserved capacity or impose restrictions topologies, e.g., existence link-disjoint routes large switch memory...
The end of Moore's law has sparked a surge on programmable accelerators (e.g., SmartNICs, TPUs) for continued scaling application performance. However, despite the great success in offloading tasks from CPU, we still lack proper mechanisms balancing load among multiple computing units present current systems. On one hand, traditional balancers (either software or hardware-based) have no visibility different server and can only dispatch requests at perserver granularity. other emerging...
We consider the problem of energy aware localized routing in ad hoc networks. In algorithms, each node forwards a message based on position information about itself, its neighbors and destination. The objective algorithms is to minimize total power for from source destination or maximize number tasks that can perform before battery depletes. this paper we extend our previous work randomized achieve high packet delivery rates show they have good overall consumption. present two different...
We present a certain approach to indoor location tracking based on sensing the strength of received RF signal. Our scheme employs database signal readings from known locations collected during prerequisite profiling stage. Subsequently, problem estimating node emitting an unknown place boils down data mining in database-to select best matched set profile points and then average coordinates those into approximate tracked sender. show preliminary experimental results which confirm our hope...
Emerging accelerated network devices, including SmartNICs (SNICs) with general-purpose CPUs - such as ARM, create new sources of queuing and processing delay in the hardware, stack, application. We therefore argue for need a framework to measure application stack performance SNIC-based host systems.