DocumentCode :
165922
Title :
An optimized RFC algorithm with incremental update
Author :
Trivedi, Uday ; Jangir, Mohan Lal
Author_Institution :
Samsung R&D Inst. India, Bangalore, India
fYear :
2014
fDate :
24-27 Sept. 2014
Firstpage :
120
Lastpage :
127
Abstract :
RFC (Recursive Flow Classification) is one of the best packet classification algorithms. However, RFC has moderate to prohibitive high preprocessing time for rule-sets having more than 10K rules. RFC does not provide incremental update. Due to these essential missing features, RFC is used in limited scenarios. This paper attempts to add these essential features in RFC. Our algorithm uses various memory and processing optimizations to speed up RFC preprocessing phase. We provide an algorithm to compute only those CBM (Class Bit Map) intersections for which corresponding value pairs are found in rules. We optimize CBM intersection by using ABV algorithm and min-max rule information. We also propose an optimized algorithm to manage real time incremental updates in RFC. The algorithm modifies only required parts of RFC tables and makes sure that the updated tables have information in correct order. For incremental update, moderate amount of extra memory is required. We tested our algorithm for preprocessing time and incremental update feature. The results indicate that we get moderate improvement in preprocessing time with real time incremental updates in our modified RFC.
Keywords :
minimax techniques; packet switching; pattern classification; ABV algorithm; CBM intersection; RFC preprocessing phase; extra memory; incremental update feature; min-max rule information; optimized RFC algorithm; optimized algorithm; packet classification algorithms; real time incremental updates; recursive flow classification; Indexes; Packet classification; RFC with Incremental update; Recursive Flow Classification;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Advances in Computing, Communications and Informatics (ICACCI, 2014 International Conference on
Conference_Location :
New Delhi
Print_ISBN :
978-1-4799-3078-4
Type :
conf
DOI :
10.1109/ICACCI.2014.6968240
Filename :
6968240
Link To Document :
بازگشت