EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

AppSlice-P4: Application-Aware In-Network Classification and Data-Plane Slicing for Heterogeneous SD-IoT Traffic

Ameer El-Sayed

Journal of Intelligent Decision Making and Information Science · 22.09.2026 · 3 (13s) · 1689-1718

DOI: 10.17762/jidmis.v3.3926 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Intelligent Decision Making and Information Science, online erschienen am 22.09.2026. Verfasst von Ameer El-Sayed. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The increasing heterogeneity of Software-Defined Internet of Things (SD-IoT) applications creates a growing need for efficient traffic differentiation and logical network slicing without excessive reliance on centralized controllers. Conventional SDN-based slicing commonly performs application identification and slice assignment at the control plane, introducing additional signaling and processing latency as network scale increases. This paper proposes AppSlice-P4, an application-aware programmable data-plane framework that integrates in-network application classification, Slice-ID assignment, and slice-aware forwarding within the P4 processing pipeline. AppSlice-P4 employs lightweight deterministic multi-field matching to distinguish Critical/Delay-Sensitive, Bandwidth-Intensive, and Best-Effort Telemetry traffic and directly maps each identified application class to its corresponding logical forwarding context. The SDN controller is retained primarily for policy management and exceptional conditions, enabling routine traffic to be processed locally. Experimental results demonstrate 99.36% classification accuracy and 99.29% slice-assignment accuracy under heterogeneous traffic. At 5,000 concurrent flows, AppSlice-P4 reduces slice-assignment latency by 95.46%, Packet-In messages by 95.73%, controller CPU utilization by 62.08%, and controller response time by 80.22% compared with controller-based slicing. The framework further achieves a 95.86% controller-offloading ratio, while maintaining 96.27% throughput efficiency, 99.08% forwarding correctness, and approximately 99.98% logical slice isolation. These results demonstrate that application-to-slice decision making can be effectively shifted into the programmable data plane, providing accurate, low-overhead, and scalable logical slicing for heterogeneous SD-IoT environments.

Zitieren

APA

El-Sayed, A. (2026). AppSlice-P4: Application-Aware In-Network Classification and Data-Plane Slicing for Heterogeneous SD-IoT Traffic. Journal of Intelligent Decision Making and Information Science, 3(13s), 1689-1718. https://doi.org/10.17762/jidmis.v3.3926

Vancouver

El-Sayed A. AppSlice-P4: Application-Aware In-Network Classification and Data-Plane Slicing for Heterogeneous SD-IoT Traffic. Journal of Intelligent Decision Making and Information Science. 2026;3(13s):1689-1718. doi:10.17762/jidmis.v3.3926

BibTeX
@article{ElSayed2026,
  author = {El-Sayed, Ameer},
  title = {AppSlice-P4: Application-Aware In-Network Classification and Data-Plane Slicing for Heterogeneous SD-IoT Traffic},
  journal = {Journal of Intelligent Decision Making and Information Science},
  year = {2026},
  volume = {3},
  number = {13s},
  pages = {1689--1718},
  doi = {10.17762/jidmis.v3.3926}
}
Kostenlos starten →