National Repository of Grey Literature 5 records found  Search took 0.00 seconds. 
Optimization of NetFlow Data Search Using nfdump
Kubovič, Martin ; Žádník, Martin (referee) ; Bartoš, Václav (advisor)
This bachelor thesis deals with optimization of NetFlow data search using the nfdump tool. This thesis describes NetFlow protocol and tool nfdump and proposes the solution using data structure Bloom filter. The main goal was to optimize data storage and processing in order to be able to search the huge amounts of collected data and get results very quickly. The outcome of this thesis is the optimized tool that network administrators can use to search these data and significantly accelerate monitoring and analyzing network.
Implementation of Bloom Filters in FPGA
Matoušek, Denis ; Kaštil, Jan (referee) ; Žádník, Martin (advisor)
The thesis deals with probabilistic data structure Bloom filter and its modifications - counting Bloom filter and multistage filter. The fundamentals of these data structures are explained and the mathematical equations describing their properties including the equation to minimize the false positive rate are shown. Appropriate hash function is chosen considering an implementation on a FPGA chip. Designs of architectures of the hash function and data structures are implemented in VHDL language and their synthesis is done. Its results are discussed considering occupied resources of a FPGA chip, a critical path and maximum frequency.
Probabilistic Packet Classification Acceleration on FPGA
Kurka, Denis ; Matoušek, Jiří (referee) ; Kekely, Lukáš (advisor)
Klasifikace síťových paketů je klíčovým úkolem v síťových systémech, protože umožňuje efektivní směrování a filtrování dat. Pravděpodobnostní filtry jsou klasifikační metoda, která používá různé techniky k aproximaci členství paketu v sadě pravidel. Tato práce zkoumá tři algoritmy: Bloomův filtr, cuckoo filtr a xor filtr. Hlavním cílem je porovnat výkon těchto tří metod při implementaci jako hardwarové komponenty v FPGA systémech. Kritéria hodnocení zahrnují chybovost, maximální frekvenci a využití zdrojů FPGA se zameřením na paměť. Výsledky ukazují, že xor filtr překonává ostatní v oblasti chybovosti, ve všech kategoriích. Bloomův filtr je nejrychlejší volbou pro menší a rychlejší komponenty, kde je vyšší chybovost tolerovatelná. Cuckoo filtr je nejefektivnější z hlediska využití FPGA logiky. Tyto poznatky přispívají k vývoji optimalizovaných klasifikačních systémů a poskytují cenné informace o možnostech implementace pravděpodobnostních filtrů v hardwarových architekturách.
Implementation of Bloom Filters in FPGA
Matoušek, Denis ; Kaštil, Jan (referee) ; Žádník, Martin (advisor)
The thesis deals with probabilistic data structure Bloom filter and its modifications - counting Bloom filter and multistage filter. The fundamentals of these data structures are explained and the mathematical equations describing their properties including the equation to minimize the false positive rate are shown. Appropriate hash function is chosen considering an implementation on a FPGA chip. Designs of architectures of the hash function and data structures are implemented in VHDL language and their synthesis is done. Its results are discussed considering occupied resources of a FPGA chip, a critical path and maximum frequency.
Optimization of NetFlow Data Search Using nfdump
Kubovič, Martin ; Žádník, Martin (referee) ; Bartoš, Václav (advisor)
This bachelor thesis deals with optimization of NetFlow data search using the nfdump tool. This thesis describes NetFlow protocol and tool nfdump and proposes the solution using data structure Bloom filter. The main goal was to optimize data storage and processing in order to be able to search the huge amounts of collected data and get results very quickly. The outcome of this thesis is the optimized tool that network administrators can use to search these data and significantly accelerate monitoring and analyzing network.

Interested in being notified about new results for this query?
Subscribe to the RSS feed.