1 / 17

Smart Pre-Classifier for Power-Efficient TCAMs in Packet Classification

This paper discusses the construction of a pre-classifier to reduce power consumption in TCAMs for multi-dimensional packet classification. The pre-classifier pre-classifies packets based on source and destination IP addresses, reducing the dimension of the problem. The rules in the classifier are reorganized and stored in different blocks, reducing the number of TCAM blocks that need to be activated for packet classification. Experimental results show significant power reductions with the proposed approach.

orozcoa
Download Presentation

Smart Pre-Classifier for Power-Efficient TCAMs in Packet Classification

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. A Smart Pre-Classifier to Reduce Power Consumption of TCAMs for Multi-dimensional Packet Classification Author: Yadi Ma, Suman Banerjee Publisher: SIGCOMM,2012 Presenter: Kai-Yang, Liu Date: 2013/1/9

  2. INTRODUCTION • For any individual packet there are usually a small number of “intercepting” rules — rules that match the given packet. • We construct a specific pre-classifier and shuffle the rules in the original classifier into different blocks so that only the pre-classifier and a small number of TCAM blocks need to be activated to classify a packet. • We reduce the dimension of the problem through the pre-classifier which pre-classifies a packet on two header fields, source and destination IP addresses.

  3. Properties of Real Classifiers

  4. Design Overview • We shuffle the rules in the classifier such that each pre-classifier entry is associated with a TCAM block, named specific block. If the number of rules that intercept with the pre-classifier entry is larger than one block size, the extra rules are stored in TCAM blocks named general blocks. • In SmartPC, each entry in a pre-classifier is constructed by expanding and combining the rectangles formed by source and destination fields of the rules in the original classifier. The rules in the classifier are reorganized, with each pre-classifier entry pointing to a TCAM block that contains all the rules covered by the corresponding pre-classifier entry.

  5. Packet Classification System for SmartPC

  6. Example (1/3)

  7. Example (2/3)

  8. Example (3/3)

  9. Packet Classification with SmartPC • With SmartPC, the actual packet classification occurs in two stages: first a given packet is classified by a pre-classifier which provides information on which TCAM block needs to be activated in the next stage; in the second stage, the block from first stage plus a few general blocks are activated and searched in parallel for a match for the packet.

  10. Summary of Real and Synthetic Classifiers

  11. The Percentage of Power Reductions

  12. The Percentage of Power Reductions

  13. The Storage Overhead of SmartPC

More Related