Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/120612
Title: High Speed Packet Classification Algorithms for Next Generation Network Services
Researcher: Agarkar B S
Guide(s): Dr Kulkarni U V
University: Swami Ramanand Teerth Marathwada University
Completed Date: 
Abstract: The routing and flow classification of IP packets is a vital component in operation newlineof Internet. The divergence and development of the Internet enforces increasing newlineexpectations from the network resources. Packet classification at link rates is vital newlinefor variety of upcoming value added next generation network services like security, newlineVoice-over Internet Protocol (VoIP), Video on Demand (VoD) and Quality of Service newline(QoS). QoS enabled routers, firewalls, network intrusion detection systems, differentiated newlineservices edge routers, etc. need to implement packet classification to find out to newlinewhich flow packet belongs to or which action is to be taken on it. Although variety newlineof hardware designs and software solutions are presented by the research community, newlinethe design and development of efficient packet classification system remains newlinea challenging problem. The factors which contribute to the difficulty are diversity newlineof rule specifications, scale of rule sets, variety of services offered, and throughput newlinerequirements of the high-speed networks. newlineThe research work presented in this thesis addresses solution to some of these newlineproblems which become a bottleneck with increasing link rates. The contributions newlineof this work include Fast Parallel Packet Classification (FPPC) algorithm, a new newlineparallel approach to packet classification to achieve high speed. Top-N selection newlinealgorithm is used to select non-overlapping active rules from the classifier according newlineto hit rate. By using these selected rules a classification data structure is constructed. newlineDisjoint grouping clustering algorithm and a new data structure are also proposed newlinewhich avoids backtracking. Hierarchical search results are passed on to the Bloom newlinefilters for final classification in place of traditional linear search. This results in newlineincreased throughput with slight decrease in memory efficiency. newlineHyper space Point Location (HPL) algorithm is a second contribution which is newlineinspired by recursive flow classification algorithm (RFCA) used for fast multidimensional packet classific
Pagination: p124
URI: http://hdl.handle.net/10603/120612
Appears in Departments:Faculty of Engineering

Files in This Item:
File Description SizeFormat 
01_title.pdfAttached File57.2 kBAdobe PDFView/Open
02_certificate.pdf57.57 kBAdobe PDFView/Open
03_abstract.pdf62.35 kBAdobe PDFView/Open
04_declaration.pdf42.2 kBAdobe PDFView/Open
05_acknowledgement.pdf44.62 kBAdobe PDFView/Open
06_contents.pdf76.03 kBAdobe PDFView/Open
07_list_of_tabels.pdf66.64 kBAdobe PDFView/Open
08_list_of_figures.pdf70.04 kBAdobe PDFView/Open
09_abbreviations.pdf62.34 kBAdobe PDFView/Open
10_chapter1.pdf955.8 kBAdobe PDFView/Open
11_chapter2.pdf2.11 MBAdobe PDFView/Open
12_chapter3.pdf1.15 MBAdobe PDFView/Open
13_chapter4.pdf1.75 MBAdobe PDFView/Open
14_conclusion.pdf68.31 kBAdobe PDFView/Open
15_bibliography.pdf90.22 kBAdobe PDFView/Open
Show full item record


Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).

Altmetric Badge: