Effects of Data Compaction and Compression on a Low-Level Communication Protocol
2026 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE credits
Student thesisAlternative title
Effekter av datakompaktering och komprimering på ett kommunikationsprotokoll på låg nivå (Swedish)
Abstract [en]
Distributed simulation systems based on the High Level Architecture (HLA) standard often include thousands of messages being sent between federates. These messages are time-sensitive and often include repetitive and redundant data. In this thesis we investigate how data compression and data compaction can be used to make the payload sizes of these messages smaller, and how that affects the total processing and transmission time of sending and receiving them. We implement a number of data compaction techniques and one data compression algorithm for a specific type of message---attribute updates---in an HLA implementation. These are then tested in terms of compression and time, using a specially developed experiment framework, and compared both to each other and a baseline. The results show that Java's built-in compression, which uses the Deflate algorithm, and the data compaction strategies outlined in this thesis---variable-byte encoding, attribute presence bits, bit packing, and retraction designator exclusion---are efficient at achieving smaller payload sizes, and in certain circumstances reduce the time. Java's compression consistently achieved the smallest payload sizes, but at the expense of increased processing and transmission time. The various data compaction techniques also successfully made the payload size smaller, with all but one achieving this without increasing processing and transmission time.
Place, publisher, year, edition, pages
2026. , p. 55
Keywords [en]
data compression, data compaction, lossless compression, communication protocol optimization, message optimization, distributed simulation, high-level architecture (HLA), runtime infrastructure (RTI), deflate, bit packing, variable-byte encoding, performance evaluation, benchmarking
National Category
Computer Engineering
Identifiers
URN: urn:nbn:se:liu:diva-226399ISRN: LIU-IDA/LITH-EX-A--26/038--SEOAI: oai:DiVA.org:liu-226399DiVA, id: diva2:2090417
External cooperation
BAE Systems OneArc Sweden AB
Subject / course
Computer Engineering
Presentation
2026-06-12, Linköping, 10:15 (English)
Supervisors
Examiners
2026-08-202026-08-062026-08-20Bibliographically approved