[show abstract][hide abstract] ABSTRACT: Nowadays, we witness a flood of continuously changing information from a variety of web sources. New challenges to track information changes in real time require new methods in web information retrieval using Multi Agents System (MAS) technology. This research continues previous work on extracting data from Online Social Networks (OSNs) by using an agent in each user profile to monitor its updates, which are sent to a controller agent that saves a history of each user‟s activity in a local repository, as well as applying a vulnerability measure to users‟ profiles. An algorithm making use of MAS within the Online Social Network Retrieval System (OSNRS) is proposed. Our experiments on data extraction show that using MAS simplifies the process of tracking profile‟s history and opens the opportunity of understanding the dynamic behavior of OSN users especially when it is combined with text mining. The application of the vulnerability measure over time highlighted that in the case of this experiment the structure of the node‟s network, rather than the contents of the node, changed over time. The validation of the vulnerability measure showed that friends of a profile, that disclose
International journal of knowledge and web intelligence. 06/2012; 3(3):p256-286.
[show abstract][hide abstract] ABSTRACT: Nowadays, we witness a flood of continuously changing information from a variety of web sources. The desire to track the information
changes in real time requires new methods in web information retrieval. This paper explores the process of retrieving information
from Online Social Networks (OSNs) using Multi Agent System (MAS) technology. The novelty of this study consists in leaving
an agent in each user profile to monitor its updates which are sent to a controller agent that saves a history of each user’s
activity in a local repository. An algorithm making use of MAS within the Online Social Network Retrieval System (OSNRS) is
proposed. Our experimental work shows that using MAS simplifies the process of tracking profile’s history and opens the opportunity
of understanding the dynamic behavior of OSN users especially when it is combined with text mining.
Agent and Multi-Agent Systems: Technologies and Applications - 5th KES International Conference, KES-AMSTA 2011, Manchester, UK, June 29 - July 1, 2011. Proceedings; 01/2011
[show abstract][hide abstract] ABSTRACT: Cloud computing adds more power to existing web cluster and internet technologies. A lot of research has been carried out to improve the security of such clouds. However, some users and organizations have not been able to exploit their benefits for several reasons: Firstly, some data is protected by privacy laws which prevent access or transport by third parties. Secondly some organisations fiercely protect their highly sensitive data behind their firewalls. Under these circumstances, the organisations need to use separate clusters for high security traffic and direct the rest of the traffic to a cloud cluster. This scenario leads to poor utilization in the local cluster which additionally increases the cost of the cloud. This paper introduces a model that solves the problem of cost minimisation by maximising local cluster utile and this is achieved by using a front-end load balancer for high and low security traffic. only when demand is high is low security traffic directed to the cloud. In this paper we analyze the model using queuing theory and -calculus. The performance metrics of throughput, delay, utilization and blocking probability were obtained via simulation under a range of parameters.
The 14th International Conference on Network-Based Information Systems, NBiS 2011, Tirana, Albania, September 7-9, 2011; 01/2011
[show abstract][hide abstract] ABSTRACT: Web portals work as a point of access to a large volume of information on the web. This paper focuses on the performance of Web portals in an E-commerce environment which involves the processing of a large number of users’ requests. It proposes a class-based priority scheme which classifies users’ requests into high and low priorities. In E-commerce, some requests (e.g. buy) are generally considered more important than others (e.g. search or browse). We contend that the requests received from a Web portal should generally get higher priority as such requests are more likely to lead to purchases. We believe that assigning such priorities at multiple service levels can improve the performance of Web portals’ requests of higher priority. The proposed scheme is formally specified and implemented, and performance results are obtained and compared to a server that does not prioritise requests. The results show significant performance improvements in the processing of high priority requests.
Journal of Systems and Software. 01/2011; 84:1373-1378.
[show abstract][hide abstract] ABSTRACT: Handover latency is the primary cause of packet loss resulting in performance degradation of the standard Mobile IPv6. Mobile IPv6 with fast Handover enables a Mobile Node (MN) to quickly detect at that IP layer it has moved to a new subnet by receiving link-related information from the link-layer; furthermore it gathers anticipative information about the new Access Point (AP) and the associated subnet prefix when the MN is still connected to the previous Corresponding Node (CN). This paper proposes an enhancement to Fast Mobile IPv6 Handover (FMIPv6), based on link layer information, we also present performance evaluations in terms of the packet loss and Handover latency using evaluation models.
Innovations in Information Technology, 2009. IIT '09. International Conference on; 01/2010
[show abstract][hide abstract] ABSTRACT: Modern E-commerce services are offered in a flexible but complex setup involving multiple websites, for example, business web portals or price comparison websites, which are used in drawing the prices of a variety of online products from related websites and presenting them to millions of customers. Though this modern style of service provisioning is very attractive it significantly increases load on the web servers implementing the E-commerce services. The concern is that overloaded servers will become unresponsive and will drop requests which are beyond their capacity. This paper presents a new approach to the synthesis of performance management mechanisms, which we contend, provides more effective way for managing the performance of modern E-commerce services. The proposed approach is formally specified and is implemented as a Java-based prototype system. A number of experiments are conducted in order to test the performance management of the proposed approach.
iiWAS'2010 - The 12th International Conference on Information Integration and Web-based Applications and Services, 8-10 November 2010, Paris, France; 01/2010
[show abstract][hide abstract] ABSTRACT: Modern E-commerce services are offered in a complex but flexible setup involving multiple websites (e.g., business web portals
or price comparison websites) with facilities for determining the quality of products. Though this modern style of service
provisioning attracts more customers it also significantly increases load on the web servers that are implementing the E-commerce
services. The concern is that overloaded servers will become unresponsive and will drop requests which are beyond their capacity.
This paper proposes a formal approach in order to investigate the effects of traffic load and the number of dropped requests
on the performance of modern E-commerce services. The proposed approach is based on a class-based priority scheme that classifies
E-commerce requests into different classes by taking into account the type of request and the client’s behaviour. The proposed
model is formally specified, implemented and tested through several experiments. The experimental results show that the proposed
approach improves the response time and throughout of high priority requests, and also analyses the consequential effect on
dropped (low priority) requests.
Analytical and Stochastic Modeling Techniques and Applications, 17th International Conference, ASMTA 2010, Cardiff, UK, June 14-16, 2010. Proceedings; 01/2010
[show abstract][hide abstract] ABSTRACT: In order to investigate the performance of priority protocols in an ecommerce system, we simulate their behaviour when composed with clients with differing behaviour. As suspected, we find that the interaction between the clients and the server is an important factor in the performance of the protocols.
Advanced Information Networking and Applications Workshops, 2009. WAINA '09. International Conference on; 06/2009
[show abstract][hide abstract] ABSTRACT: Among the characteristics of the future wireless networks is the desire to support a wide range of wireless users and a diverse set of services from many different types of networks. A unifying technology for diverse wireless networking is mobile IP, a protocol established by the Internet Engineering Task Force (IETF) that allows mobile nodes (MN) to change their point of attachment to the Internet while still being able to maintain a connection to the network mobile IP is designed to support the mobility of a user, it is the only current means for offering seamless roaming to mobile computers in the Internet. In this paper, we give a brief survey of mobile IP operation and functions, as well as proposed pi-calculus algorithm of enhancement mobile IP (E-mobile IP) handoff mechanism using link layer information to enhance the performance in term of packet loss and handoff latency.
[show abstract][hide abstract] ABSTRACT: Q2 abstract This paper is an investigation into the performance of E-commerce applications. E-commerce has become one of the most popular applications of the web as a large population of web users is now benefiting from various on-line services including product searches, product purchases and product comparison. E-commerce provides users with 24-7 shopping facilities. However, the consequence of these benefits and facilities is the excessive load on E-commerce web servers and the performance degradation of E-commerce (eCom) requests they process. This paper addresses this issue and proposes a class-based priority scheme which classifies eCom requests into high and low priority requests. In E- commerce, some requests (e.g. payment) are generally considered more important than others (e.g. search or browse). We believe that by assigning class-based priorities at multiple service levels, E- commerce web servers can perform better and can improve the performance of high priority eCom requests. In this paper, we formally specify and implement the proposed scheme and evaluate its performance using multiple servers. Experimental results demonstrate that the proposed scheme significantly improves the performance of high priority eCom requests. & 2008 Published by Elsevier Ltd.
J. Network and Computer Applications. 01/2009; 32:455-460.