Projects

A Dynamically Scaled Cloud-based Web-Service using Docker-Swarm, HAProxy and OpenStack

Screen Shot 2017-08-02 at 19.22.27.jpg

In this report, a cluster of cloud-based web-service is built in order to adjust the number of Docker-based web-servers (containers) in order to manage the incoming rate of user requests. The main focus of this cloud-based web-services is to prevent service depletion due to heavy traffic from outside network and automate the nodes to generate new services to adapt the requests approaching for accessing those web-services. To achieve this infrastructure, several tools such as Docker-Swarm, Docker-Machine, HAProxy have been used. Additionally, Grafana, InfluxDB and cAdvisor have been used to monitor the traffic pattern as well as CPU, memory and file-system usage of the nodes. Upon successful testing with the ApacheBench to exhaust the traffic handling capability of the load-balancer, the services are auto-scaled in order to adjust the concurrent connection requests and to stabilize the network overload.

Link to PDF: Project2

Standard
Projects

An Automated Test Framework for Web-Pages using Git, Jenkins, Docker and OpenStack

 

technical_designIn this report, an automated continuous integration suite is created in order to achieve a test framework for web-pages using Git, Jenkins, Docker and OpenStack. The main focus of this framework is the automation of the entire process. Using this framework the web-developers can observe if the newly modified HTML codes (which are pushed from Git-repository) are working successfully or not before it is released to the production. Upon successful test build, the web- server is automatically running in to production which ensures the complete automated continuous integration test suite for web-pages.

 

Link to PDF: Project1

Standard
Projects

Thesis: Long Distance, Low Data Rate Packet Transmission at 868 MHz: Protocol Design and Implementation

Data transmission over wireless networks can be achieved in various ways. Some of these technologies are feasible in many real life scenarios, but there are many consequences we have to consider in this field, such as packet loss, quality degradation of radio signals, ambient noise, transmission delay and etc. For the low data rate packet transmission, to adapt and to address IoT is one of the most discussed topics nowadays, especially in terms of transmission protocol design and the implementation of that design in real life scenarios. Transmitting packets in low data rate over long distance efficiently is becoming an important issue in this fast-paced modern world. Our objective is to achieve long distance with low data rate packet transmission by combining them in the same protocol design with diverse form of services.
In this project, we design an embedded solution and implement our own transmission schemes based on Raspberry Pi micro-controller to send low date rate packets via radio modules that operating on 868 MHz band to achieve long distance. We embed and implement the design with low cost and low power components and send real-time services such as GPS co-ordinates, temperature, image and etc. over wireless network. We develop our design to relay the transmission from the source node to the destination node by using a forwarding node in between them to achieve further distance. The designed schemes responds to each other with acknowledgements assuring the successful transactions between the nodes and for the request of retransmissions. The schemes are developed further, where the retransmission will occur with an adaptive data rate at the end of the transmission process to ensure the successful transfer of a large file. The data are collected and examined on the sender and receiver ends and calculate parameters such as packet loss rate, transmission delay, etc., by running various tests in different scenarios and optimising the experimental results.

 

Link to PDF: Master Thesis – Spring 2016

Standard
Projects

Energy Harvesting-aware Backoff Algorithms for Distributed Device-to-Device Communication

Due to limited energy capacity in mobile devices, energy harvesting is regarded as a promising technology for lifetime extension in device-to-device (D2D) communication. However, medium access control (MAC) protocols designed for D2D communication intrinsically do not consider the existence of harvested energy. In this paper we study four backoff (BO) algorithms, among which one algorithm is proposed to enhance the BO selection based on energy harvesting rate and residual energy of devices. Simulations are performed in order to assess how the proposed algorithm can extend the lifetime of a D2D communication network as well as to provide fair channel access opportunities among devices.

DOI: 10.1109/VTCSpring.2016.7504226

IEEE VTC 2016-Spring, 186-49514, 15-18 May 2016

Standard
Projects

Voice Transmission over Low Data Rate Wireless Networks: Embedded Systems and Implementation

Voice transmission over wireless networks can be done in various ways. Some of these technologies are feasible in many real life scenarios, but there are many consequences we have to consider in this field, such as packet loss and quality of the voice degradation. These proposed solutions are mainly focused on the voice transmission over high data rate network, considering the higher level of data rate on those transmissions. But for the low data rate voice transmission, not much work has been done. especially in terms of practical design and the implementation of that design. In this project, we design an embedded solution and implement our design to transfer voice data over low data rate enabled components. We embed and implement the design with low cost and low power components and send voice/audio data over a personal area network (PAN). At the end, we collect the data from our embedded design by running the tests and optimising the numerical results. The numerical results show that with this design the voice data can be sent over low data rate network in real time with very little latency and relatively long range in distance.

Standard
Projects

MAC Enhancement and Performance Evaluation in Wireless Sensor Networks with Energy Harvesting

Wireless sensor nodes are usually battery equipped and designed to maximise the energy conservation in order to avoid as much as battery extortion. Considering the limited amount of energy available in sensor nodes, proper distribution of harvested energy is appear to be most promising technology in wireless sensor networks (WSN). Harvested energy arrival rate cannot be predicted in advance due to its stochastic behaviour. This demands the necessity of designing an efficient MAC protocol for WSNs. In this project, we design four energy harvesting aware backoff (BO) algorithm for CSMA/CA and duty cycle scheme to extend the lifetime of sensor nodes by administering energy harvesting rate as well as residual energy. We enhanced backoff algorithm with energy harvesting wireless nodes using CSMA/CA MAC mechanism to extend the lifetime of overall sensor networks. In addition, we evaluate the duty cycle mechanism by comparing lifetime of three different schemes considering energy harvesting rate as a control parameter. The simulation results shows that MAC mechanism with energy harvesting awareness enhance the performance of WSNs. Our numerical results shows EH and residual energy based backoff algorithm for CSMA/CA enhances the lifetime of WSNs.

Standard