Proposals for MSc Thesis are provided to students from any country. Just send an email to get more information about a specific proposal. You have to ask the preliminary approval to your local supervisor, such that your work will be co-supervised.
Proposed and supervised by Prof. Ilias Savvas (Dept. of Computer Science and Engineering, University of Applied Sciences of Thessaly — TEI of Thessaly), Larissa, Greece.
Title: Increasing the performance of data clustering techniques
Download the proposal: PDF
ComplexSim is a C-based simulation platform intended to support the study of P2P systems and complex networks.
ComplexSim architecture is based on two layers, the parallel simulation kernel, which manages the execution of simulations on SMP systems, by means of an API enabling the scheduling of tasks and events, and the complex network data and runtime.
Experimental results have shown that, even in the case of huge complex networks with hundred million nodes, ComplexSim exhibits better performances, in terms of memory consumption and processing times, with respect to similar solutions.
The only limitation of ComplexSim is the system memory, which is saturated as the network size reaches the order of billions. A simple, lightweight and fast DBMS (nosql) may be integrated to avoid this limitation.
The main assignment is to integrate the fast DB originally written by Google, LevelDB (http://code.google.com/p/leveldb/).
C/C ++ programming skills are required.
JarvSis is a distributed scheduler capable to automate the execution of multiple heterogeneous tasks on IoT applications by means of a modular and adaptable software architecture. The design of JarvSis is designed to nteract with any devices, even simple sensors, that expose heterogeneous remote interfaces, e.g. web-api, MQTT message capabilities, or ROS (Robotic Operating System).
Through JarvSis, workflows can be composed by a multitude of heterogeneous tasks, and a JarvSis network is constructed by allocating the control agents from the Cloud to the Fog, in order to control the IoT device running in the “ground”.
- JarvSis to ROS. By means of the rosbridge suite ( http://wiki.ros.org/rosbridge_suite ), we aim to create a module offering a flexible interface that allows JarvSis to invoke tasks of ROS devices.
- JarvSis to MQTT. We aim to implement the module that allows JarvSis to invoke tasks of any devices having the capabilities to communicate by means of MQTT. To this end, the Paho implementation of the MQTT protocol should be used (https://eclipse.org/paho/ ).