Národní úložiště šedé literatury Nalezeno 1 záznamů.  Hledání trvalo 0.00 vteřin. 
Probabilistic Packet Classification Acceleration on FPGA
Kurka, Denis ; Matoušek, Jiří (oponent) ; Kekely, Lukáš (vedoucí práce)
Classifying network packets is a crucial task in networking systems, as it allows for efficient routing and filtering of data. Probabilistic filters are a classification method that uses different techniques to approximate the membership of a packet in a set of rules. This work investigates three algorithms: Bloom, cuckoo, and xor filter. The main aim is to compare the performance of these three methods when implemented as hardware components in FPGA systems. The evaluation criteria include error rate, maximal frequency, and FPGA resource usage, primarily focusing on memory. The results indicate that the xor filter outperforms the others regarding error rate, which is superior in any error rate category. The Bloom filter is the fastest option for smaller and quicker components where a higher error rate is tolerable. The cuckoo filter is the most resource-efficient when FPGA logic is the primary concern. These findings contribute to the development of optimised classification systems and provide valuable insights into the possibilities of implementing probabilistic filters in hardware architectures.

Viz též: podobná jména autorů
1 Kurka, Daniel
Chcete být upozorněni, pokud se objeví nové záznamy odpovídající tomuto dotazu?
Přihlásit se k odběru RSS.