Seminar Topics

Topics of Jakob Mass

Monitoring & Analysing Mobility in Smart Cities

The Smart City concept involves gathering large amounts of data from different sensor sources (Vehicle counters, public transport, cell tower data, etc), the challenge lies in fusing the different data to value. A theme of interest here is analysing, predicting how the inhabitants of a city move about, as this helps make better decisions to avoid traffic congestion, etc.

• RQ: How to manage the fact that sensors cannot observe 100% of mobility events in a city?
• RQ: How far ahead can predicitons about mobility be made, what influences this?
• RQ: How to determine the "weak link" data sources in sensor fusion?
• BuScope: Fusing Individual & Aggregated Mobility Behavior for "Live" Smart City Services
• [El-Tawab'20] A Framework for Transit Monitoring System Using IoT Technology: Two Case Studies
• RQ: Does context detection and monitoring conflict with Android frameworks energy-saving & security policies?
• [Chatterjee'20] Detecting Mobility Context over Smartphones using Typing and Smartphone Engagement Patterns
• SimRa: Using crowdsourcing to identify near miss hotspots in bicycle traffic

Mobile Computing

Mobiles in Edge Computing: Code Offloading and Applications

A requirement for systems which enable offloading of mobile device code/tasks to other (non-mobile) devices such as fog nodes is an execution environment that matches the original runtime (e.g. Android runtime). A typical solution is to use virtual machines or containers, but in case of mobile such as Android or iOS, these approaches are not always feasible. Another challenge is deciding what to offload, when and to which machine. This topic explores the possible approaches and system designs of the runtime/environment for code offloading in Android.

• RQ: What limitations are imposed on offloaded code and why? I.e. what kind of code can/can't be offloaded.
• Android Unikernel: Gearing mobile code offloading towards edge computing
• RQ: How to model & predict execution time of offloaded tasks?
• RQ: What is the impact of the phones network (cellular, Wi-Fi) on task offloading?
• ULOOF: A User Level Online Offloading Framework for Mobile Edge Computing

Infrastructure and System Design for Edge & IoT

IoT System Development Framework

Designing and implementing IoT systems can be complex, as it involves activities at the cloud-level (storage, analysis, processing tasks), client-level (applications and user-facing services) and device-level (hardware, sensor/actuator device software). Usually, the technologies for these 3 levels are significantly different with little overlap. Some frameworks aim to unify this picture and aim to give a single set of tools, programming languages for designing IoT systems at all the levels.

• RQ: What kind of use cases benefit from using a single unified language for the 3 IoT solution parts (device side, the cloud side, and the client side), and in what are possible drawbacks for other kinds of applications?
• TinyLink 2.0: integrating device, cloud, and client development for IoT applications

Applications for Mobile, IoT, Edge & Miscellanous

Computation Offloading for Vehicles

Modern vehicles are equipped with computing and communication modules, that enable connecting with each other and road-side infrastructure. This has lead to research in various ways how computational tasks can be distributed between nodes in a vehicular scenario. This topic tries to survey the field where Fog/Edge Computing and Vehicles meet, and identify applications which benefit more from using the Fog/Edge approach.

• RQ: What types of offloading are prevalent for Vehicular scenarios?
• RQ: What characterizes applications used in vehicular computational offloading?
• Computation Offloading for Vehicular Environments: A Survey

Identifying Business Process activities from IoT Sensor Data [TAKEN]

IoT device data enable the opportunity of fine-grained monitoring, analysis of business process. IoT can also helps detect when some-sub activity of a process is started or ended (e.g. washing machine has finished wash cycle and now starts drying cycle). But in practice, reliably detecting and segmenting IoT data to corresponding process steps is not easy. This topic studies the literature to find what are the state of the art methods for discovering activities from raw IoT sensor data (supervised, unsupervised machine learning, visualization-based)

• RQ: How can sensor data be associated with process activity executions to potentially automate the detection of activities from sensor data?
• Method to Identify Process Activities by Visualizing Sensor Events
• Supervisor: Chinmaya

Smart Networking, Security for IoT

Backscatter networking for Internet of Things

In backscatter radio communication, a device transmits data to a receiver by modulating and reflecting a signal from a third, external signal source, as opposed to generating the signal itself and sending it directly to the receiver. This approach is useful mainly due to the decreased energy requirements as the transmitter does not have to generate a radio signal itself, only remodulate it. As a result, this technology is also interesting for IoT devices. This topic studies the existing prototypes &solutions of backscatter networking for IoT Backscatter radio communication exploits the reflectedor backscattered signals to transmit data, where thebackscattered signals can be the reflection of ambientradio frequency (RF) signals, the RF signals from thededicated carrier emitter

• RQ: What are the limitations and drawbacks of using back-scatter networking for IoT?
• Gateway over the air: towards pervasive internet connectivity for commodity IoT

(more backscatter papers in mobisys'20 conference)

Topics of Chinmaya Dehury

email: chinmaya.dehury@ut.ee

C1: A short review on clustering algorithms for edge computing environment

Edge computing refers to bringing the computing capacity to the edge of the network. The edge devices are equipped with a limited computing capacity. When , we talk about edge computing, we need think about the characteristics of edge data, edge resource, edge devices and many more. On the other hand clustering strategies in IoT and wireless communications are applied mostly to reduce the energy consumption and further improve the network resource utilization. In this topic, student needs to go through the algorithms/strategies that specifically designed for cluestering the edge data or the devices or the users.

• Clustering Algorithms on Low-Power and High-Performance Devices for Edge Computing Environments, https://www.mdpi.com/1424-8220/21/16/5395
• Task Classification and Scheduling Based on K-Means Clustering for Edge Computing, https://doi.org/10.1007/s11277-020-07343-w
• A Comprehensive Survey of Clustering Algorithms, https://doi.org/10.1007/s40745-015-0040-1

C2: X Discovery: A short survey

Discoverability mechanism let an entity X on a network be discoverable. The enetity X can be a service, a device, a network or knowledge. For instance, the service discovery protocol is used by a client device to find out about the services it can use on a server device. On the other hand, edge computing allows the IoT device generated data or client data to be processed by the nearest computing device present at the periphery of the network. This topic would focus on study of discovery system in the edge computing, especially the study of device, service, and knowledge discovery at the edge. Some of the questions the student need to focus on, are:

• What do you mean by X discoverability?
• How the X discovery works at the edge, where X can be device, service, or knowledge?

Some of the references to start with:

• Device Discovery in D2D Communication: A Survey (https://ieeexplore.ieee.org/abstract/document/8835011)
• Service Discovery (https://www.dfki.de/~klusch/i2s/SD_essay_klusch2013.pdf)
• Towards Service Discovery and Invocation in Data-Centric Edge Networks (https://ieeexplore.ieee.org/abstract/document/8888081)
• Collaborative Learning-Based Industrial IoT API Recommendation for Software-Defined Devices: The Implicit Knowledge Discovery Perspective (https://ieeexplore.ieee.org/abstract/document/9208715)

C3: Intelligence discoverability and obeservability in edge infrastructure

Discoverability is the degree to which something, especially a piece of content or information, can be found in a search of a file, database, or other information system. - Wikipedia
Observability is the ability to measure the internal states of a system by examining its outputs. - Splunk
This topic would focus on the Discoverability and Observability aspects of edge computing infrastructure (including edge intelligence, edge knowledge, IoT data, edge services, edge devices, etc). Some of the precise questions, the student needs to focus are:

• Q: What is the difference between Discoverability and Observability in general?
• Q: What can be made be discoverable and observable in edge computing?
• Q: How do you see these two term in the context of Edge computing?

Some resources to start with:

• Discovery systems in ubiquitous computing (https://ieeexplore-ieee-org.ezproxy.utlib.ut.ee/document/1626210)
• Edge-to-Edge Resource Discovery using Metadata Replication (https://zenodo.org/record/5851025/files/27.pdf)

C4: Mapping Dogont ontology with smart home data

DogOnt ontology aims at offering a uniform, extensible model for several IoT devices, e.g. temperature sensors, CO2 sensor, and many more. The ontology allows you to describe the device location, device capabilities, interface, technology-specific features, etc. OWLready2 is a Python library for managing the classes, subclasses, entities, instance, and properties of dogont ontology. This topic focuses on creating the data models that are aligned with the dogont ontology.

Some sources:

• DogOnt - Ontology Modeling for Intelligent Domotic Environments, DOI: 10.1007/978-3-540-88564-1_51, http://iot-ontologies.github.io/dogont/
• An ontology design pattern for IoT device tagging systems, https://ieeexplore.ieee.org/document/7356558

C5: Modelling languages for smart environments

DogOnt ontology aims at offering a uniform, extensible model for several IoT devices, e.g. temperature sensors, CO2 sensor, and many more. Similarly, DTDL from Azure, allows you to create, modify and visualize the smart city or smart building data such as City objects, parking, environment, ports, streetlights, people, places etc. This topic would focus on looking at different modelling languages (similar to Dogont or DTDL) available for smart city.

• Q- What are the modeling languages available for smart city?
• Q- What are the advantages and limitations of those modelling languages?

Ref:

• https://github.com/Azure/opendigitaltwins-dtdl
• https://github.com/Azure/opendigitaltwins-smartcities
• https://docs.microsoft.com/en-us/azure/digital-twins/overview

Your topic

You may come with your own topic.....

Topics of Shivananda R Poojara

(ALREADY TAKEN) Topic 1: Predictive maintenance of IoT devices System monitoring is a common activity to ensure good quality of service in applications deployed in edge and fog environments. However due to stochastic workloads and the deployed environment of devices may be prone to outages, faults, and errors(e.g., CPU or HDD faults). Here, once fault generated propagates over the system leading to performance degradation of the application and system. Insensitive IoT applications such as Healthcare and Industrial applications, a small failure is not desirable otherwise can lead to catastrophic failure. So, it's important to monitor the devices by logging the data and predicting the failures in advance. Predictive maintenance is a technique of monitoring and predicting the failures in a system. Memory card faults or persistent failures are a major problem in IoT devices. There are several factors that implicitly affect the failures of the SD card such as the deployment of devices in harsh environments, development of bad blocks and malware attacks, etc. The following would be research questions that need to address:

• RQ1 : What are different techniques used in predictive maintenance?
• RQ2 : Predictive maintenance techniques for IoT devices
• RQ3 : Howd PD techniques applied for predicting SD Card failures?

Related research articles to start with:

 Distributed Continuous-Time Fault Estimation Control for Multiple Devices in IoT Networks

(ALREADY TAKEN) Topic 2: Intelligence at the extreme edge using TinyML

Rapid growth in ICT accelerated the large-scale deployment of IoT applications ubiquitously in several domains, including smart factories. Here, more often, the requirement is to process the data near to source device for faster decisions. The idea is to leverage the nearby servers to run Machine Learning algorithms. But, these algorithms require huge memory and CPU to train and run, but this is inevitable in resource-contained devices. On the other side, a set of techniques and tools helps to build memory and CPU-efficient ML models that can fit various embedded system architectures. one such platform is tinyML. So, the topic aimthe s to investigate the tinyML framework or tool to build edge-centric ML applications.

• RQ1 : What is TinyML framework and its architecture?
• RQ2 : What is Reformable TinyML?
• RQ3 : Taxonomy of Reformable TinyML?

Related research articles to start with: Intelligence at the Extreme Edge: A Survey on Reformable TinyML TinyML foundation A rev.iew on TinyML: State-of-the-art and proan spects

(ALREADY TAKEN) Topic 3: Performance modeling of serverless computing

Analytical models are proven to be an efficient way of estimating performance and designing new systems, including cloud computing. Not only for performance tuning but also used to estimate the cost incurred in cloud environments. Many research works focused on, the performance and cost modeling of virtual machines in cloud systems. The goal of the study is to list and describe the analytical methods used to tune the mean response time, utilization, and throughput in serverless workloads.

• RQ1 : What are the different analytic approaches used to model serverless platforms?
• RQ2 : How does analytical modeling improve the quality of service and resource utilization and reduce the operational cost of serverless platforms?

Related research articles to start with: Performance Modeling of Serverless Computing Platforms

Topic 4: Estimating the optimal size of serverless functions Serverless is new cloud computing model for delivering the run time services as service based function level billing. It has advantages of granular scaling of fine grained servcies (functions) and helps for optimized billing. However, most of the pdevelopers ,vendors a focus on handling resource management tasks such as resource provisioning, deployment, and auto-scaling. Setting up of configeration of the serveless function left to developers and its very crucial task becuase that can lead to inconsistent costs in the system.

• RQ1 : How to decide the resource configuration of the serverless functions?
• RQ2 : How to estimate the function size without prior performance tests?

Related research articles to start with: Sizeless: predicting the optimal size of serverless functions

Topic 5: Serverless Simulators The main goal of this topic is to list and describe the various FaaS simulators. The FaaS Simulators are one way of estimating the prior cost and performance metrics without running on real serverless cloud platforms, that helps the developers to test and tune the workloads.

• RQ1 : Why FaaS simulatirs are necessary?
• RQ2 : List and explain the architecture of various FaaS Simulators available?

Related research articles to start with: SimFaaS: A Performance Simulator for Serverless Computing Platforms

Some of the other topics would be:

• Digital twin model for IoT applications
• COSCO: Edge and Cloud Computing Coupled Simulator
• Is edge computing reality or hype?
• ekuiper: Lightweight IoT data streaming analytics engine for edge computing

Topics of Pelle Jakovits

(ALREADY TAKEN) Topic P.1. Monitoring the quality of Internet of Things time series data and data aquisition integrations.

• RQ1 : Can IoT data integration and the quality of metrics be monitored without compromising privacy (looking at raw data)?
• RQ2 : What are the most improtant metrics to measure to estimate the quality of IoT data?

Related research articles to start with:

• G-O. Meritxell, B. Sierra and S. Ferreiro, "On the Evaluation, Management and Improvement of Data Quality in Streaming Time Series," in IEEE Access, vol. 10, pp. 81458-81475, 2022, doi: 10.1109/ACCESS.2022.3195338. https://ieeexplore.ieee.org/abstract/document/9845398

Topic P.2. Traceable data sharing in IoT and Smart Cities

• RQ1 : How to trace the origin of IoT data to be able to prove when and where exactly data was produced?
• RQ2 : How to trace the origin of things (in IoT) and produce moving along supply chains using IoT technologies and data?

Related research articles to start with:

• Li, Chunpei, et al. "Efficient and traceable data sharing for the Internet of Things in smart cities." Computers and Electrical Engineering 103 (2022): 108389. https://www.sciencedirect.com/science/article/pii/S0045790622006061

(ALREADY TAKEN) Topic P.3: Comparison of software platforms (like Apache NiFi) that are used to automate the data aquisition, migration and transformation in-route.

• RQ1 : What are the most improtant characteristics of tools that are used for building big data aquisition pipelines and data integrations services?
• RQ2 : What are the respective differences, andvantages and disadvantages of major data quisition platforms?

Related research articles to start with:

• Rooney, Sean, et al. "Experiences with managing data ingestion into a corporate datalake." 2019 IEEE 5th International Conference on Collaboration and Internet Computing (CIC). IEEE, 2019.
• Hamadou, Hamdi Ben, Torben Bach Pedersen, and Christian Thomsen. "The Danish National Energy Data Lake: requirements, technical architecture, and tool selection." 2020 IEEE International Conference on Big Data (Big Data). IEEE, 2020. https://vbn.aau.dk/ws/portalfiles/portal/400693632/FEDDL_submited_version_Copyright.pdf

(Topic is TAKEN) Topic P.4: Practical Edge Mesh

Traditional approaches for the Internet of Things (IoT) data processing have focused on collecting data originating from different devices in private data centers or in clouds and processing it centrally. Full dependence on remote cloud services can lead to the lack of autonomy of applications in case of network degradation or failures. For example, Smart city applications that depend on logic being executed in remote clouds will not be resilient in accident and disaster scenarios when cloud services may not be accessible anymore. This has driven the implementation of Edge and Fog computing platforms, where some data processing tasks or other cloud services are moved closer to the data sources to reduce data sent to the cloud and to improve data privacy. The goal of this topic is to investigate a new type of Edge computing model called Edge Mesh and study what benefits it brings in comparison to cloud-centric data collection and processing, and typical Edge computing approaches.

• RQ1 : What is Edge Mesh, how does it differ from usual approaches to Edge computing and Fog computing?
• RQ2 : How is Edge Mesh different from Mist computing?
• RQ2 : What are its benefits in comparison to cloud-centric data collection and processing, and Edge computing.

Related research articles to start with:

• On the Road from Edge Computing to the Edge Mesh

Topic P.5: Open Source Fog computing platforms: Eclipse ioFog

Traditional approaches for Internet of Things data processing have focused on collecting data originating from different devices in private data centers or in clouds and processing it centrally. Full dependence on remote cloud services can lead to the lack of autonomy of applications in case of network degradation or failures. For example, Smart city applications that depend on logic being executed in remote clouds will not be resilient in accident and disaster scenarios when cloud services may not be accessible anymore. This has driven the implementation of Edge and Fog computing platforms, where some data processing tasks or other cloud services are moved closer to the data sources to reduce data sent to the cloud and to improve data privacy. The goal of this topic is to investigate existing open-source Edge and Fog computing frameworks and to give an overview of their architectures, involved challenges, and the current state-of-the-art in the field. The student should particularly focus on Eclipse ioFog as the reference framework.

• RQ1 : What are the main barriers in industry adopting the Fog and Edge computing models in comparison to using cloud-centric approach?
• RQ2 : What are the main aspects of Eclipse ioFog, what makes it different from previous apoproaches for designing Fog computing platforms?
• RQ3 : Are platofms like Eclipse ioFog already used in large scale industry?

Related research articles to start with:

• Toward a Reference Architecture Model for Industrial Edge Computing

(Topic is TAKEN) Topic P.6: Cloud deployment modeling for data science and ML pipelines

Cloud deployment models have been used to describe how the cloud systems and services should be deployed together with software, infrastructure, network, and any other resources which are needed. TOSCA is one open standard language for such modelling, and is used in platforms like Cloudify. Cloud providers have their own custom models and frameworks for modeling and enacting such deployments. Furthermore, DevOps and Infrastructure as Code (IaS) methods have been applied to speed up such deployments, for example Ansible, Terraform, etc. The goal of this topic is to study the deployment modeling of ML pipelines and data science applications and pipelines in general.

• RQ1 : Are current Deployment modelling (e.g. TOSCA, AWS solutions) sufficient to model the deployment and dynamic execution of ML and data science pipelines
• RQ12 : What (if any) extensions are required to TOSCA to support modelling the deployment and dynamic execution of ML and data science pipelines

Related research articles to start with:

• Chinmaya Kumar Dehury, Pelle Jakovits, Satish Narayana Srirama, Giorgos Giotis, Gaurav Garg, TOSCAdata: Modeling data pipeline applications in TOSCA, Journal of Systems and Software, Volume 186, 2022, 111164, ISSN 0164-1212, https://doi.org/10.1016/j.jss.2021.111164.
• Falkenthal, M., Breitenbücher, U., Képes, K., Leymann, F., Zimmermann, M., Christ, M., ... & Kempa-Liehr, A. W. (2016, November). Opentosca for the 4th industrial revolution: Automating the provisioning of analytics tools based on apache flink. In Proceedings of the 6th International Conference on the Internet of Things (pp. 179-180).

Topic P.7: Sensor Data fusion in Internet of Things and Smart Cities

A large number of sensors have and are being deployed in IoT networks and Smart Cities. However, it may not be feasible to deploy a new sensor every time a new type of observation/data needs to be detected. It is sometimes more feasible to use existing sensors and data to compute or extrapolate new observations - by performing data fusion. The goal of this topic is to study and give an overview of Data Fusion, how it is being applied in these domains and provide an overview of challenges and research gaps related to them.

Related research articles to start with:

• Billy Pik Lik Lau, Sumudu Hasala Marakkalage, Yuren Zhou, Naveed Ul Hassan, Chau Yuen, Meng Zhang, U-Xuan Tan, A survey of data fusion in smart city applications, Information Fusion, Volume 52, 2019, Pages 357-374, ISSN 1566-2535, https://doi.org/10.1016/j.inffus.2019.05.004

(Topic is TAKEN) Topic P.8: State-of-the-art solutions for Multi-organisational Smart city Data Marketplaces

The ability to share data between different organizations is necessary for applications and services to take full advantage of IoT networks, devices, and data. Building Smart Cities requires integrating, analyzing, and sharing data generated in the city belonging to a large number of organizations in the city. There have been quite a few attempts to create marketplaces where companies, data providers, and users can share and even sell their data but there are not many existing successful platforms. The goal of this topic is to investigate what are the current state-of-the art data marketplaces in this domain and also give an overview of previous attempts. Investigate what are the main challenges in creating data marketplaces for sharing real-time data streams.

• RQ1 : How ready are the existing platforms for supporting large-scale smart city scenarios?
• RQ2 : How well is real-time-streaming data supported?

Related research articles to start with:

• Spiekermann, Markus. "Data marketplaces: Trends and monetization of data goods." Intereconomics 54.4 (2019): 208-216.
• Koutroumpis, Pantelis, Aija Leiponen, and Llewellyn DW Thomas. The (unfulfilled) potential of data marketplaces. No. 53. ETLA Working Papers, 2017.

Or propose your own ideas

Some examples:

• Comparison of State-of-the-Art (SOTA) open-source IoT platforms, focusing on platforms that provide features for both device integration and data storage. The theoretical part involves a survey into the most popular SOTA platforms. The practical part involves setting up and demonstrating a selected candidate platform.
• Designing digital twins for Smart City visualization and monitoring
• Synthetic data generators for large-scale testing of IoT and Smart City systems