No abstract

No abstract

No abstract

<span>Developments in computer networking have raised concerns of the associated Botnets threat to the Internet security. Botnet is an inter-connected computers or nodes that infected with malicious software and being controlled as a group without any permission of the computer’s owner. <br /> This paper explores how network traffic characterising can be used for identification of botnet at local networks. To analyse the characteristic, behaviour or pattern of the botnet in the network traffic, a proper network analysing tools is needed. Several network analysis tools available today are used for the analysis process of the network traffic. In the analysis phase, <br /> the botnet detection strategy based on the signature and DNS anomaly approach are selected to identify the behaviour and the characteristic of the botnet. In anomaly approach most of the behavioural and characteristic identification of the botnet is done by comparing between the normal and anomalous traffic. The main focus of the network analysis is studied on UDP protocol network traffic. Based on the analysis of the network traffic, <br /> the following anomalies are identified, anomalous DNS packet request, <br /> the NetBIOS attack, anomalous DNS MX query, DNS amplification attack and UDP flood attack. This study, identify significant Botnet characteristic in local network traffic for UDP network as additional approach for Botnet detection mechanism.</span>

Semantic Web approach with the assistance of ontology is widely used to give more reliable application in retrieving information and knowledge.  It is capable to discover the World Wide Web (WWW) that is presented in natural-language text.  Based on previous research, incorporating categorization with ontology concept has proven to give better results.  However, performing hybrid of the search engine using another technique that is user profiling has a promising potency in enhancing the searching process.  Utilizing searching time and giving relevant results are the contributions of this research.  The proposed hybrid techniques integrate ontologies, categorization and user profiling concept.  In user profiling, similarity measure is adopted in making comparison between two different ontologies.  WordNet and UTHM Onto are the independent ontologies used in this process.  The preliminary experimental results have given interesting results in terms of data arrangement and time usage.