Accelerating Packet Classification via Direct Dependent Rules

Fuchino, Takashi; Harada, Takashi; Tanaka, Ken

doi:10.1109/nof52522.2021.9609820

Cited by 3 publications

(2 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the reordering method using the weight calculated before reordered can not account for the rule that will have a bigger weight by weight fluctuation. Thus, many of previous methods that use weights can not reduce efficiently the latency [1]- [6].…”

Section: The Problem Of Previous Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Computational Complexity of Allow Rule Ordering and Its Greedy Algorithm

Fuchino

Harada

TANAKA

et al. 2023

IEICE Trans. Fundamentals

View full text Add to dashboard Cite

Packet classification is used to determine the behavior of incoming packets in network devices according to defined rules. As it is achieved using a linear search on a classification rule list, a large number of rules will lead to longer communication latency. To solve this, the problem of finding the order of rules minimizing the latency has been studied. Misherghi et al. and Harada et al. have proposed a problem that relaxes to policy-based constraints. In this paper, we show that the Relaxed Optimal Rule Ordering(RORO) for the allowlist is NP-hard, and by reducing from this we show that RORO for the general rule list is NP-hard. We also propose a heuristic algorithm based on the greedy method for an allowlist. Furthermore, we demonstrate the effectiveness of our method using ClassBench, which is a benchmark for packet classification algorithms.

show abstract

Section: The Problem Of Previous Methodsmentioning

confidence: 99%

“…In many cases, placing rules that match a large number of packets at the top of the list reduces the latency. Therefore, the problem of finding an order of rules that minimizes the latency, the optimal rule ordering problem (ORO), has been studied [1]- [6]. This problem is known as NP-hard [7].…”

Section: Introductionmentioning

confidence: 99%