Research On Optimization Technology Of Open Flow Flow Table In Software Defined Networks

Software Defined Network(SDN)decouples the control plane and the data plane,provides the network with open and flexible programming ability,enhances the management capabilities of the network,improves the flexibility and scalability of network function,and paves the way for future network architecture research.The expansion of supported fields in OpenFlow improves the fine-grained control ability of Open Flow in data forwarding plane,but the arbitrary wildcard support in match field relies heavily on TCAM for matching process,which brings additional pressure to the storage space of TCAM(ternary content addressable memory)in physical device.In addition,the explosion of network traffic requires a large number of flow table rules to deal with the data packets,but the storage space for flow table in OpenFlow switch is limited.Therefore,how to optimize the OpenFlow flow table from the perspective of reducing the storage overhead and ensuring the functionality of the flow table is an urgent problem to be solved.The optimization of route table,access control list,etc in traditional physical devices mainly use compression.Given the compression method in traditional network,we focus on the unique characteristics of the match field in OpenFlow flow table and study 3 aspects in the following 2 types.The first type is to study the “horizontal” compression method to decrease the bit-width of table entries and reduce the redundancy between match fields in flow table.The second one is to study “vertical” compression method to reduce the number of table entries without modifying forwarding semantics.1.A function-integral TCAM-saving “horizontal” compression method for flow tableA Function-Integral TCAM-saving Compression model for flow table of OpenFlow(FICO)is proposed to keep the fine-grained control ability of Open Flow in the “horizontal” compression method.FICO analyzes the relationships among all the fields and then categorizes the redundancy among different fields into three types.Based on the three redundancy types,three compression algorithms named as inter-field merge,field mapping and intra-field compression are presented,which are mainly performed in RAM space.The outcomes of each compression algorithm are flow entries with smaller bit-width which is sent to TCAM for flow matching.In this way,the flexibility of Open Flow is not harmed,thus maintaining the function integrity of the original flow table.Simulation shows that the compression performance of FICO improves about 5.3% with no damage to the function integrity of OpenFlow,and the compression performance stands stable with the increase of flow table size.2.A “vertical” compression method for application-aware routingA dynamic flow aggregation method in SDN for application-aware routing is proposed to improve the average compression ratio and reduce the average flow table aggregation time.We first utilize instruction slicing to produce an instruction-independent partition,design and then apply the entry merging algorithm to merge flow entries in each subset of a partition.Comparing with the existing method,the average compression ratio(the average compression time)of the proposed method is better(smaller).3.A table entry merging and placement metnod based on function compositionA function composition method of SDN based on topology transformation is proposed to reduce the storage overhead of Open Flow switch,decrease the method complexity and consider the function bearing capacity of the single switch.We first propose the overall function composition framework,model function composition problem as a two-stage mapping process of “composition strategy-execution path-flow table entry”,and then put forward a method of table entry merging and placement.When single switch can not support needed function,we propose a topology transformation method to scatter the entry composition in multiple nodes and carry out the allocation and placement of entries.The experimental results show that the proposed method can effectively reduce the storage cost of the entries in TCAM,improve the efficiency of the composition as well as reduce the processing delay,and meet the needs of multiple function processing for data packets.