Posts by Collection



Collaborative computing of parallel applications in P2P environments

Published in NEA 2015, 2015


This book addresses the issue of “How to explore collaborative computing in P2P Grids to run BSP applications efficiently?”. To answer this question came the BSPonP2P model, which creates an environment with approaches based on the structured and unstructured models from the P2P networks. The combination of these models aims to streamline the management of communication and data on the network. Another difference of the proposed model is the use of the parallel programming model Bulk Synchronous Parallel (BSP), which creates an environment for executing processes, validating dependencies and improving communication between them. We use metrics required for the processes as memory, computing, communication, and also information from the infrastructure to establish a unique performance metric called PM. This metric is evaluated periodically to define migrations according to the launched superstep barriers. The results show that BSPonP2P uses the computational idleness, and brings performance gains to the executions.

Recommended citation: da Silva Veith, Alexandre; Righi, Rodrigo. R

Pain-o-vision, effortless pain management

Published in MobiSys '21: Proceedings of the 19th Annual International Conference on Mobile Systems, Applications, and Services, 2021

[Paper] [BIBTEX]

Recommended citation: Ramprasad, Brian; Chen, Hongkai; da Silva Veith, Alexandre; Truong, Khai; de Lara, Eyal.



CR05: Advanced Topics in Scalable Data Management

Graduate course, ENS-Lyon, Computer Science, 2018

General Presentation

Big data is becoming a bigger deal every day. Every day 2.5 Quintillion bytes of data are created. In 2016, big data will affect everyone. These lectures focus on a large point of view of the current data management. In charged of Lecture 8 (Data Stream Processing) and lecture 9 (Practical Session of Data Stream Processing).

CSC2228: Advanced Topics in Mobile and Pervasive Computing: Edge Computing

Graduate course, University of Toronto, Computer Science, 2021

General Presentation

Next generation mobile and IoT applications, such as intelligent personal assistants, medical patient monitoring, and intelligent transportation systems, are not compatible with the existing cloud computing model where applications are deployed on a small number of extremely large datacenters accessible over the wide area network. These applications require either low latencies that are not achievable over the wide area, or produce large volumes of data that can overwhelm the network. Edge computing is a new paradigm that addresses these issues by expanding the traditional cloud architecture with additional datacenter layers that provide computation and storage closer to the end user. For example, a wide-area cloud datacenter which serves a large country can be augmented by a hierarchy of datacenters that provides coverage at the city, neighborhood, and building level. This course provides an overview of some of the current research directions in edge computing..