在SDN 中壓縮三元表達式之轉發表

Abstract

軟體定義網路(SDN)，近年來蓬勃發展，許多關於其效率的研究及論文已被發表。依 照需求動態調整網路已變得十分容易基於控制層與資料層的分離。TCAM 是OpenFlow 交換器中用來儲存轉發表之儲存元件，其具有平行處理之能力讓他在封包轉發、繞送 查詢、封包分類有著較高的效率，但其也有著造價昂貴、耗能較大和產生較大個熱能 須發散等問題，使得提高TCAM 的效能是有必要的。 轉發表在資料流通過交換器時扮演了極為重要的角色，它決定該資料流的該如何行 動。而轉發表是儲存於TCAM 中，因為前面所說的理由，所以我們想藉由提高轉發表 所能處理的資料流量來減少因溢出而頻繁地更新轉發表進而降低更動TCAM 的次數來 達到提升TCAM 的效率。 我們利用了三元字串表達式和貪心法概念的優勢，提出了一個壓縮轉發表的方法。 我們的方法在壓縮率及壓縮時間上有著卓越的表現。比較於其他兩種壓縮ACL 的壓縮 方法，我們在壓縮時間上有著大幅的領先，而在壓縮率上也與其中一最佳化壓縮方法 有著不相上下的結果。在SDN 環境中，我們的方法是比較適合且有著較佳的結果。

Software Defined Network(SDN), a new type of network are developing fast these years, more and more thesis and paper are proposed on issue of efficiency and cost. Because of the separation between control plane and data plane, modify network configuration dynamically now being easier.TCAM is the storing unit of OpenFlow switch which store flow table for forwarding flows. But with these special hardware, some problem show up like power consumption, generates a high level of heat that must be dissipated and expensive to build, so to improve efficiency of TCAM is necessary. Flow table play an important role when flows pass though the switches, it decide how these flows perform. Flow table is storing in TCAM, so we trying to improve TCAM’s efficiency by increase flow table’s utilization rate. Compressing flow table is a way to reduce number of rules to keep flow table without frequently update because of flow table overflow. We take advantage of ternary string expression and greedy method and come up with an approach to compressing flow table. Our method have an excellent result of compression time and compression ratio. Compare to two ACL compressor, our approach leading on compression time and have as good as an optimal solution on compression ratio. Our method is more suitable for SDN environment without interrupting controller and have excellent result.