SARA -- A Semantic Access Point Resource Allocation Service for Heterogenous Wireless Networks

TL;DR

The paper addresses AP selection in heterogeneous wireless networks by proposing SARA, an ontology-driven autonomic resource allocation service built on the SEANET knowledge-based management system. SARA reasons over a semantic, ontology-backed knowledge base to choose among technologies like LTE, LiFi, WiFi, and Satellite, incorporating real-time metrics and user-defined rules (e.g., via first-order logic) to adapt AP assignment. The authors validate SARA in Mininet-WiFi, demonstrating substantial throughput improvements over traditional SSS approaches, with most stations more than doubling their throughput and some achieving multi-fold gains, while maintaining a responsive overhead of around $3$ seconds per re-selection. The results support the viability of semantic, rule-driven AP selection for dynamic, multi-technology networks and highlight the practical benefits of integrating knowledge-based management with autonomic network control.

Abstract

In this paper, we present SARA, a Semantic Access point Resource Allocation service for heterogenous wireless networks with various wireless access technologies existing together. By automatically reasoning on the knowledge base of the full system provided by a knowledge based autonomic network management system -- SEANET, SARA selects the access point providing the best quality of service among the different access technologies. Based on an ontology assisted knowledge based system SEANET, SARA can also adapt the access point selection strategy according to customer defined rules automatically. Results of our evaluation based on emulated networks with hybrid access technologies and various scales show that SARA is able to improve the channel condition, in terms of throughput, evidently. Comparisons with current AP selection algorithms demonstrate that SARA outperforms the existing AP selection algorithms. The overhead in terms of time expense is reasonable and is shown to be faster than traditional access point selection approaches.

Figures (10)

• Figure 1: Architecture of the SEANET, with the proposed components: knowledge base generator, SPARQL engine and a network management API over the heterogenous networks.
• Figure 2: Knowledge entities about Access Point in the SEANET knowledge base.
• Figure 3: The part of the ToCo Ontology used in SARA.
• Figure 4: The movement of 30 mobile stations at different time point.
• Figure 5: Result end-to-end throughput of stations which the SARA algorithm is applied. The stations shown in the figure are Sta 1, 2, 3, 4, 6, 7, 10, 12, 15, 16, 17, 18, 19, 20, 21, 26. The red dash line in each figure denotes the time when SARA algorithm is executed.