EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

HiCache-P4: Intelligent Rule-Placement Decisions Through Hierarchical In-Switch Caching for Scalable Policy Enforcement in SD-IoT Networks

Ameer El-Sayed

Journal of Intelligent Decision Making and Information Science · 01.10.2026 · 3 (15s) · 440-465

DOI: 10.17762/jidmis.v3.4525 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Intelligent Decision Making and Information Science, online erschienen am 01.10.2026. Verfasst von Ameer El-Sayed. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The rapid expansion of software-defined Internet of Things (SD-IoT) networks introduces significant challenges for intelligent rule management and scalable policy enforcement, particularly when dynamic flow arrivals and large rule populations exceed the limited TCAM capacity of programmable switches. Conventional single-level caching mechanisms often evict reusable rules, increasing cache misses and dependence on centralized controller decisions. Addressing this problem requires efficient decisions about which rules should occupy high-speed memory, which should remain available in secondary storage, and when controller assistance is necessary. This paper proposes HiCache-P4, an intelligent hierarchical rule-caching framework that integrates an L1 TCAM Hot Rule Cache with an L2 SRAM/register-based Warm Rule Store. The framework employs a lightweight, activity-aware decision mechanism based on rule-access frequency and recency to guide rule placement, promotion, demotion, replacement, and expiration. Its sequential lookup process prioritizes fast L1 resolution, enables local recovery from L2, and invokes the SDN controller only after a complete local miss. HiCache-P4 is implemented in a P4-based SD-IoT environment and evaluated against TCAM-only, LRU, and LFU caching approaches under varying flow churn, rule-access locality, and rule-table pressure. The results demonstrate total local hit ratios of 96.5%, 92.0%, and 84.0% under low, moderate, and high churn, respectively. Under high churn, L2 recovers an additional 19.0% of lookup requests, reducing controller-assisted lookup demand by 52.9% relative to LFU. For mixed hot/cold traffic, HiCache-P4 achieves a 92.0% total hit ratio, compared with 79.0% for LFU, highlighting the value of activity-aware decisions that preserve intermittently reused rules. With a fixed 256-entry L1 cache and a 1,024-entry L2 store, the framework maintains 1,210 locally resolvable rules at a generated population of 4,000, representing a 4.73× increase in effective local rule capacity over single-level caching without enlarging the fast-cache allocation. These findings demonstrate that lightweight, access-driven rule-placement decisions can improve local policy availability, reduce controller dependency, and support scalable resource management in dynamic SD-IoT environments.

Zitieren

APA

El-Sayed, A. (2026). HiCache-P4: Intelligent Rule-Placement Decisions Through Hierarchical In-Switch Caching for Scalable Policy Enforcement in SD-IoT Networks. Journal of Intelligent Decision Making and Information Science, 3(15s), 440-465. https://doi.org/10.17762/jidmis.v3.4525

Vancouver

El-Sayed A. HiCache-P4: Intelligent Rule-Placement Decisions Through Hierarchical In-Switch Caching for Scalable Policy Enforcement in SD-IoT Networks. Journal of Intelligent Decision Making and Information Science. 2026;3(15s):440-465. doi:10.17762/jidmis.v3.4525

BibTeX
@article{ElSayed2026,
  author = {El-Sayed, Ameer},
  title = {HiCache-P4: Intelligent Rule-Placement Decisions Through Hierarchical In-Switch Caching for Scalable Policy Enforcement in SD-IoT Networks},
  journal = {Journal of Intelligent Decision Making and Information Science},
  year = {2026},
  volume = {3},
  number = {15s},
  pages = {440--465},
  doi = {10.17762/jidmis.v3.4525}
}
Kostenlos starten →