軟體定義網路中藉由建立初始路由規則樹有效率減少路由規則使用之演算法

An Algorithm of Construction of Default Rule Tree for Efficient Utilization of Routing Table in Software-Defined Network

張益銘(Chang, Yi Ming)

指導教授 ： 蔡明哲

國立清華大學/電機資訊學院/資訊工程學系/碩士(2015年)

https://doi.org/10.6843/NTHU.2015.00496

摘要

由於在軟體定義網路中路由規則表是個相當珍貴的資源,導致如何解省其使用是許
多人研究的方向。 文章當中,我們藉由建立一棵新的路由規則樹取代舊的最短路
徑路由規則樹來幫助減少我們設計一個減少路由規則表使用的路由演算法。 我們
演算法針對的對象是在資料中心中小流量的資料傳遞 (佔總體資料流比例85%),並
且可將釋放出來的路由表空間提供給大流量資料流使用,以提升整體網路效能。
我們設計此樹的設計理念是能夠讓此樹當成高速公路做路由規劃, 使用此樹所走得
路徑不需要增加多的路由規則來傳遞封包。 最後既然大家都能藉由此樹來做封包
傳遞, 也有可能造成其傳遞路徑的變長,因此最後我們將增加使用者的路由長度限
制。
關鍵字:路由規則樹, 有效率路由規劃

關鍵字

路由規則樹 ； 有效率路由規劃

並列摘要

Efficient utilization of routing table is an major issue in software-defined network(SDN)
due to the high price of the lookup routing table. In this paper, we focus on mini-
mizing the usage of TCAM of mice flows (85 percentage overall) in the data center.
The liberated spaces of the routing table by the algorithm could be use for increasing
the throughput of the network by other traffic engineering functions. First of all, we
provide a routing policy is to reuse the default rule as many as possible in order to
decrease the TCAM table size. Next, different from the original default tree, which
is built as a shortest path tree, we construct a new default rule tree called CDRT
tree which could efficiently reuse by the routing policy. Further more, the tree is con-
structed by using linear optimization model. At last, we evaluate the performance of
CDRT using Bcube topology, and compare with two routing policy. One is shortest
path routing, which is the original routing policy for mice flows, the other is to reuse
shortest path tree for routing. At last, avoiding the routing path overstretch because
of the reuse of the routing method, we would set length limitation on the routing
path.
Keywords: Default Rule Tree, Efficient Routing.

並列關鍵字

Default Rule Tree ； Efficient Routing

參考文獻

[1] B. experience with a globally-deployed software defined wan, “Sushant jain,” in

[2] X.-N. Nguyen, D. Saucez, C. Barakat, and T. Turletti, “Optimizing rules place-

ment in openflow networks: trading routing for better efficiency,” in ACM

[3] N. Bitar, S. Gringeri, and T. J.Xia, “Technologies and protocols for data center

and cloud networking,” IEEE Communications Magazine, vol. 51, no. 9, pp.