Power-Saving Hybrid CAMs for Parallel IP lookups

IP lookup with the longest prefix match is a core function of Internet routers. Partitioned Ternary Content Addressable Memory (TCAM)-based search engines have been widely used for parallel lookups despite power inefficiency. In this paper, to achieve a higher throughput and power-efficient IP lookup, we introduce hybrid CAM (HCAM) architecture with SRAM. In our approach, we break a prefix into a collapsed prefix in CAM and a stride in SRAM. This prefix collapse reduces the number of prefixes that results in reduced memory usage by a factor of 2.8. High throughput is achieved by storing the collapsed prefixes in partitioned CAMs. Our record for 2 BGP tables shows that an HCAM saves 3.6 times energy consumption compared to the TCAM based implementation.

[1]  Bin Liu,et al.  An ultra high throughput and power efficient TCAM-based IP lookup engine , 2004, IEEE INFOCOM 2004.

[2]  Haoyu Song,et al.  Packet classification using coarse-grained tuple spaces , 2006, 2006 Symposium on Architecture For Networking And Communications Systems.

[3]  Patrick Crowley,et al.  CAMP: fast and efficient IP lookup architecture , 2006, ANCS '06.

[4]  Srihari Cadambi,et al.  Chisel: A Storage-efficient, Collision-free Hash-based Network Processing Architecture , 2006, 33rd International Symposium on Computer Architecture (ISCA'06).

[5]  Mehrdad Nourani,et al.  A TCAM-Based Parallel Architecture for High-Speed Packet Forwarding , 2007, IEEE Transactions on Computers.

[6]  Viktor K. Prasanna,et al.  Beyond TCAMs: An SRAM-Based Parallel Multi-Pipeline Architecture for Terabit IP Lookup , 2008, IEEE INFOCOM 2008 - The 27th Conference on Computer Communications.