OpenFlow 網路中改良式單鏈路故障恢復之機制

Abstract

隨著網際網路持續的成長與越來越複雜化，若在訊息傳遞時發生故障，即使是短時間的故障也會對通訊造成巨大的影響。因此再故障發生前先做好相關恢復機制以減少故障發生後所造成的影響就顯得相當重要，在OpenFlow網路中也不例外。 本論文探討在OpenFlow網路中改良式單鏈路故障恢復之機制，提出了兩種改良模型，且在各種不同的網路環境下模擬，從模擬結果可得知，改良後的兩種單鏈路故障恢復機制之模型不僅僅能夠在故障發生當下先經由次要路徑轉送受影響的封包至目的地，且在交換器傳送PORT_STATUS訊息給控制器後，在模型一中控制器修改最佳且合適的次要路徑之優先權，將其規劃為故障後的備用主要路徑；模型二中控制器將已計算好的備用主要路徑其流程條目(Flow entry)安置在路徑上所經過的交換器其流程表(Flow table)中，兩種模型皆能節省傳輸成本及增加網路可靠度。

With the internet continued growth and become more and more complicated, if a single link failure occurs when the message was delivered, even is a short-term failure that can also cause a significant impact on communication. Therefore, it is very important to pre-plan a related recovery mechanisms in order to reduce the impact caused by the link fault, and it's no exception in the OpenFlow networks. This thesis investigates the improved single link failure recovery mechanisms in OpenFlow networks, we propose two improved models and simulate in different network environments, from the simulation results that show the two improved single link failure recovery mechanisms can forward the affect packets to destination immediately when the failure occurs. After the switch sent the PORT_STATUS message to the controller, in the system model 1, the controller modify the better and appropriate secondary path priority as the backup primary path for link broke; in the system model 2, the controller install the backup primary path that pre-computed flow entries in the traversed switch flow table, two system models are enhance the transmission efficiency and increase network reliability