Slot-Size Adaptation and Utility-Based Packet Aggregation for IEEE 802.15.4e Time-Slotted Communication Networks

Time-slotted communication is used in countless protocols and systems. IEEE 802.15.4e time-slotted channel hopping (TSCH) is one of those examples which has shown remarkable performances in the literature. However, time-slotted systems have one fundamental drawback: a slot is predefined to be suffic...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE internet of things journal 2024-05, Vol.11 (9), p.16382-16397
Hauptverfasser: Kim, Hongchan, Lee, Geonhee, Shin, Juhun, Paek, Jeongyeup, Bahk, Saewoong
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Time-slotted communication is used in countless protocols and systems. IEEE 802.15.4e time-slotted channel hopping (TSCH) is one of those examples which has shown remarkable performances in the literature. However, time-slotted systems have one fundamental drawback: a slot is predefined to be sufficiently long enough to accommodate one exchange of a maximum-sized packet and an acknowledgment. If most packets in the system are far smaller than the maximum, a significant amount of residue time within each slot is wasted, leading to corresponding loss in effective data rate. To address this fundamental challenge, we propose utility-based adaptation of slot-size and aggregation of packets (ASAP) which reduces wasted time in slotted systems to improve throughput and latency. ASAP consists of two orthogonal approaches: 1) slot-length adaptation (SLA) dynamically adapts timeslot length to actual packet size distribution and 2) utility-based packet aggregation (UPA) transmits aggregated packets in multiple consecutive slots to maximize slot utility. We case-study ASAP in the context of TSCH. We implement ASAP on real embedded devices and evaluate on large-scale testbeds using state-of-the-art schedulers to demonstrate a 2.21\times improvement in throughput as well as a 78.7% reduction in latency.
ISSN:2327-4662
2327-4662
DOI:10.1109/JIOT.2024.3354051