Research And Design Of Hybrid TCAM Based On Nonvolatile Memory

Software-Defined Network(SDN)is an emerging network paradigm that simplifies network management by decoupling the control plane and data plane,where switches act as data forwarding devices and network management is controlled by logically centralized servers.By shifting the control plane to a logically centralizedcontroller,SDN offers programmable functions to dynamically control and manage packets forwarding and processing in switches.In SDN,network flow is managed by a set of associated rules that are maintained by switches in their local Ternary Content Addressable Memories(TCAMs).A TCAM can compare a packet at line rate with all the patterns containing in the installed rules at the same time.However,commodity switches support relatively few rules,in thousands or tens of thousands.On the other hand,fine-grained rich SDN policies result in tremendous amounts of rules,which need to be stored and maintained in SDN switches.While TCAMs become indispensable due to their performance benefits,it comes with non-negligible cost.TCAM is power hungry,expensive and takes up quite a bit of silicon space.The TCAM in modern commercial switch is based on SRAM.The high-power leakage and low density of SRAM restrict the TCAM capacity significantly.The performance of packet processing is severely constrained by the capacity of TCAM,which aggravates the processing burden and latency issues.As the data amount generated from Internet keeps growing,to improve the TCAM capacity while minimizing the TCAM usage cost becomes a challenge.Recently,the non-volatile memory(NVM)such as resistive RAM(ReRAM)has been introduced to implement TCAM due to its merits of low leakage power and high density.Compared with the conventional sTCAM,NVM-based TCAM(nvTCAM)is reported to achieve higher storage capacity and lower energy consumption.However,nvTCAM is not perfect due to its inherent high latency of write operation.Furthermore,the update operation towards the installed rules in TCAM is a slow procedure.Today's SRAM-based switches only support around 40 to 50 rule-table updates per second.The update performance will be more serious for nvTCAM-based switches,highly constraining the network performance in large networks.In this paper,we propose a hybrid TCAM architecture which consists of NVM-based TCAM(nvTCAM)and SRAM-based TCAM(sTCAM),utilizing nvTCAM to cache the most popular rules to improve cache-hit-ratio while relying on a very small-size sTCAM to handle cache-miss traffic to effectively decrease update latency.Considering the special rule dependency,we present an efficient Rule Migration Replacement(RMR)policy to make full utilization of both nvTCAM and sTCAM to obtain better performance.Experimental results show that the proposed architecture outperforms current TCAM architectures.