authors: shuai ding, zhen chen, and zhi liu publisher: icndc 2012 presenter: chai-yi chu date:...

Post on 12-Jan-2016

218 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

1

Parallelizing FIB Lookup in Content Centric Networking

Authors: Shuai Ding, Zhen Chen, and Zhi Liu

Publisher: ICNDC 2012

Presenter: Chai-Yi Chu

Date: 2013/03/20

2

Introduction Implementation◦Generating FIB and Interest◦ Parallelizing FIB Lookup

Experiment

Outline

3

Leverages multi-core platform to accelerate the FIB lookup in CCN router.

Based on TILEPro64 platform, which has 64 identical tiles.

Two parallelized lookup algorithms1. based on hash table. 2. based on Bloom filter

Introduction

4

5

Use a special server called ASN server, which returns corresponding AS number upon receiving a query of domain name.

6

7

8

TILEPro64 multicore platform◦ 64 identical tiles, each of which is a full featured processor.◦ 43 of them are available to user space programs.

FIB Generator◦ generates FIB entries.

Interest Generator ◦ generates Interests to search in FIB.

FIB ◦ implements lookup algorithms.

Implementation

9

10

Simulate CCN Interests with http URL requests.◦ extract 50,000 URLs from realistic pcap files captured at the

gateway of an office. Generate FIB ◦we set a seed for a random number generator and use statistics

of domain names. Generate Interest

1. generate a FIB prefix using the same seed as generating FIB.

2. produce the suffix using another random number generator based on the statistics of URL path.

Generating FIB and Interest

11

Hash table based lookup algorithm Start by the longest prefix, each prefix is searched in

FIB until a certain prefix matching is found.

Parallelizing FIB Lookup

12

13

Bloom filter is searched at first, if a prefix doesn’t exist in Bloom filter, there is no need to search for it in the hash table any more.

14

15

load factor is 1.0

Experiment

16

load factor is 10.0

17

Hash based

top related