软件定义网络中可扩展的流表项处理机制  被引量:2

Scalable Flow Table Entries Processing Mechanism in Software-defined Networks

在线阅读下载全文

作  者:朱向阳[1] 陈兵[1] 

机构地区:[1]南京航空航天大学计算机科学与技术学院,江苏南京211106

出  处:《计算机技术与发展》2016年第12期12-16,共5页Computer Technology and Development

基  金:江苏省科技项目(未来网络前瞻性研究项目)(BY2013095-2-10)

摘  要:数据中心内部的流量特性和OpenFlow交换机流表的超时机制及其容量的限制为软件定义网络(SDN)的应用带来了可扩展性难题,即由于老鼠流频繁调用控制器以及流表超时机制带来的重复路由请求,导致控制器平面和OpenFlow信道成为SDN架构的性能瓶颈而无法根据业务的需求提供有QoS保证的服务。针对此问题从两个角度提出了改进措施。为交换机配置以目的地址为导向的高优先级的通配符流表项,使得大部分老鼠流能够被转发平面直接处理而不需要询问控制器;在控制器和转发平面之间增加一个动态索引哈希缓存层,当流表项因超时或者流表空间耗尽被流表丢弃时将其缓存在缓存层中,此时收到重复的路由请求时能够避免重复计算。实验结果表明,提出的通配符流表项能够有效降低网络的延迟和流表项装载数量,而基于动态索引哈希的缓存层有着较高的命中率和查找速度。The time, out mechanism and capacity limitation of flow table in OpenFlow switches, together with the traffic characteristics in data centers,have brought scalability problems in the application of Software-Defined Network (SDN). Due to the frequent and repeatable routing request invocations by mice flows and flow table timeouts, the SDN control plane and OpenFlow secure channel will become the performance bottleneck of whole SDN architecture, and thus cannot provide QoS guaranteed services according to network resources and business requests. Two improvement measures are put forward from two aspects. The switches are configured with destination address oriented wildcard flow table entries for mice flows, making that most of mice flows are handled by forwarding plane directly without ask- ing control plane. A dynamic-index hash cache layer is added between the control plane and the forwarding plane,and flow table entries will be put into the cache layer when they are evicted from flow table because of timeout or other reasons, which can avoid repeatable routing calculations. Experimental results show that wildcard flow table entries can effectively reduce the network delay and the number of flow table entries, and the cache strategy proposed has higher hit rate and looking up speed.

关 键 词:软件定义网络 老鼠流 流表超时 通配符流表项 动态索引哈希 

分 类 号:TP311[自动化与计算机技术—计算机软件与理论]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象