2A Internship: Simulating IoT Data Exchange using Queueing Theory
Developing Internet of Things (IoT) applications requires to take into consideration software and hardware components traversing different layers of the OSI model. In particular, specific APIs and protocols must be chosen to enable data exchange between sensors and actuators. Then, forwarding and processing software components must be configured to process raw data. Finally, proper hardware nodes and access networks must be selected to support the applications’ requirements (range, network capacity, etc). The above choices significantly affect the resulting performance of IoT applications. IoT developers require tools that are capable to rapidly estimate multiple Quality of Service (QoS) metrics (latency, system utilization, memory, etc) of IoT applications, depending on the selected software and hardware components.
This project aims to introduce a simulation tool for rapid analysis of data exchange interactions in the IoT by relying on Queuing Theory. The selected team will have to leverage a existing open source queueing simulators (JMT, developed by Politecnico di Milano and Imperial College London or JINQS, developed by the Imperial College London) to implement extensions related to the characteristics of IoT systems.
The selected candidate must get familiar with the JMT and JINQS simulation tools and then extend one of the two simulators with the following features:
- lifetime validity of customers entering a queue (it models obsolete IoT data);
- intermittently available (on-off) queueing servers (it models mobile IoT data producers/consumers);
- variable service rate of queueing servers (it models different data transmission rates);
- probabilistic dropping at the input of queues (it models dropping of IoT data based on, e.g., the data importance);
- message prioritization for different classes of customers (it models prioritization of different groups categorized by IoT data sizes, available network bandwidth and data importance);
- heterogeneous arrival rates of customers (it models realistic arrival rates of IoT applications);
After implementing the above extensions to existing queues, the main objective will be to enable the composition of existing/extended queues as queueing networks for representing the performance of IoT applications.
Last year of engineering school.
Skills & qualities
- Fluent in English
- Knowledge of Object Oriented Programming.
- Knowledge of Java.
- Knowledge of Java Collection framework.
- Ability to analyze existing software code base.
- Familiarity with development tools for managing and building software projects (Eclipse) and version control systems (Git).
- Starting date: June 2021
- Duration:: 2 months
- Location: Telecom SudParis, Evry (https://www.telecom-sudparis.eu/en/)
Georgios Bouloukakis, georgios.bouloukakis AT telecom-sudparis.eu