Donta, P. K. (2023, July 21). Unleashing the Potential of Distributed Computing Continuum Systems [Conference Presentation]. 2nd China-Austria Innovation Cooperation Exchange Conference & International Symposium on AI and Digital Economy (CAICC-IDIA 2023), Nanjing, Jiangsu Povince, China.
2nd China-Austria Innovation Cooperation Exchange Conference & International Symposium on AI and Digital Economy (CAICC-IDIA 2023)
en
Event date:
20-Jul-2023 - 21-Jul-2023
-
Event place:
Nanjing, Jiangsu Povince, China
-
Keywords:
Distributed Computing Continuum Systems; Representation learning; Intelligent Data Protocols
en
Abstract:
A distributed compute continuum system (DCCS) makes use of interconnected devices to harness collective computing power. As part of this work, unlearning is categorized into monitoring and governing devices, representation learning is used to extract insights, and intelligent data protocols are developed for efficient data exchanges. Effective monitoring ensures that devices operate effectively, or detects potential bottlenecks, which are addressed through governing models. Representation learning can extract meaningful insights from all DCCS data sources, enabling more efficient knowledge extraction and decision-making systems. Further, intelligent data protocols within DCCS facilitate efficient data communication, coordination, and synchronization among the participating devices according to the system's condition. These protocols not only make the system more autonomous but also address several challenges including congestion control, dynamic message expiry, prioritization and message filtering in different data protocols.
