Publications 2022

[j22.1]
S. Bosse, , PSciLab: An Unified Distributed and Parallel Software Framework for Data Analysis, Simulation and Machine Learning—Design Practice, Software Architecture, and User Experience , Appl. Sci. 2022, 12(6), 2887; 10.3390/app12062887
Paper PDF Publisher
A hybrid distributed-parallel cluster software framework for heterogeneous computer networks is introduced, which supports simulation, data analysis, and Machine Learning (ML) using widely available JavaScript Virtual Machines (VM) and Web browsers to perform the working load. This work addresses parallelism primarily on control-path level and partially on data-path level targeting different classes of numerical problems that can be either data-partitioned or replicated. composed from a set of interacting worker processes that can be easily parallelised or distributed, e.g., for large-scale multi-element simulation or ML. The suitability and scalability for static- and dynamic sized problems is experimentally investigated with respect to the proposed multi-process and communication architecture and the data management using customised SQL data bases with network access. The framework consists of a set of tools and libraries, mainly the WorkBook (processed by a Web Browser) and the WorkShell (processed by node.js). It can be shown that the proposed distributed-parallel multi-process approach with a dedicated set of inter-process communication methods (message- and shared-memory-based) scales efficiently with the problem size and the number of processes. Finally, it is shown that the JavaScript-based approach can be easily used for exploiting parallelism by a typical numerical programmer and data analyst not requiring any special knowledge about parallel and distributed systems and their interaction. There is a focus on VM processing.

Book Cover

[r22.1]
Stefan Bosse, BeeTS: Smart Distributed Sensor Tuple Spaces combined with Agents using Bluetooth and IP Broadcasting, CoRR abs/2204.02464 (2022)
Paper PDF
Most Internet-of-Things (IoT) devices and smart sensors are connected via the Internet using IP communication driectly accessed by a server that collect sensor information periodically or event-based. Although, Internet access is widely available, there are places that are not covered and WLAN and mobile cell communication requires a descent amount of power not always available. Finally, the spatial context (the environment in which the sensor or devices is situated) is not considered (or lost) by Internet connectivity. In this work, smart devices communicate connectionless and ad-hoc by using low-energy Bluetooth broadcasting available in any smartphone and in most embedded computers, e.g., the Raspberry PI devices. Bi-directional connectionless communication is established via the advertisements and scanning modes. The communication nodes can exchange data via functional tuples using a tuple space service on each node. Tuple space access is performed by simple evenat-based agents. Mobile devices act as tuple carriers that can carry tuples between different locations. Additionally, UDP-based Intranet communication can be used to access tuple spaces on a wider spatial range. The Bluetooth Low Energy Tuple Space (BeeTS) service enables opportunistic, ad-hoc and loosely coupled device communication with a spatial context.
[b22.1]
S. Bosse, L. Dahlhaus, U. Engel, Web data mining: collecting textual data from web pages using R, in: Handbook of computational social science ; Volume 2: Data science, statistical modelling, and machine learning methods, Editors U. Engel et al., 2022, Routledge ISBN 9781032111391 DOI
Chapter PDF Publisher
[b22.2]
S. Bosse, Large-scale agent-based simulation and crowd sensing with mobile agents, in: Handbook of computational social science ; Volume 2: Data science, statistical modelling, and machine learning methods, Editors U. Engel et al., 2022, Routledge ISBN 9781032111391 DOI
Chapter PDF Publisher