ATEO: Adaptive Timeout & Eviction Orchestrator
Flow table management has a huge role in improving scalability and efficiency in Software-Defined Networking (SDN). In this paper, adaptive timeout assignment and eviction orchestrator (ATEO) are presented, which provides timeout assignment with Z-score-based classification, reactive heuristic-based eviction, and dynamic threshold adjustments. ATEO dynamically changes timeout values based on the flow’s characteristic, assuring that frequently used flows and elephant flows stay in the table while others are removed efficiently. Its heuristic-based eviction strategy detects and removes the flows expected to be the least needed to free up space. Additionally, dynamic threshold adjustments allow ATEO to respond to different network conditions in real time. Therefore, excessive controller interactions and a higher number of rejected flows can be prevented. ATEO’s simulation results show a significant leap compared to existing methods in minimizing packet-in while achieving competitive flow table utilization compared to its competitors.
Authors
- Beytüllah Yiğit (ORCID: https://orcid.org/0000-0002-6700-5012)
- Fatih Alagöz (ORCID: https://orcid.org/0000-0002-2189-5371)
- Emre Ertekin
Institutions
- Boğaziçi University (TR)
Publication Details
- Published
- 2025-06-23
- DOI
- https://doi.org/10.1109/blackseacom65655.2025.11193954
- Primary Topic
- Real-Time Systems Scheduling
- Type
- article
- Field-Weighted Citation Impact
- 0.00