Computer Science

Permanent URI for this collection

https://hdl.handle.net/10530/366

Browse

Now showing 1 - 20 of 62

Quality of service provisioning in mobile wireless network
(2001) Nkambule, Mafika William.; Adigun, M.O.; Ojong, G.
One major challenge in the design of wireless mobile networks is the provision of quality of service (QoS) guarantees. Mobile users would require resources to be available in the cells they would move Into in order to maintain the same QoS as in the current cell. However, reservation in all neighbouring cells or based on the user's mobility history results in over-reservation of resources, leading to schemes that are very inefficient and unreliable. This dissertation presents a next-cell prediction scheme (NCPS) that uses a sectored-cell approach to predict a next cell, a mobile that will move into with an open connection. The NCPS was simulated and its performance was evaluated under different speeds of the mobile users. The behaviour of the scheme was also observed under different sizes of the critical region. The results of the scheme were also compared to the results of a representative history-based scheme. It was found that the NCPS performs better than the history-based schemes. The simulation results show that prediction accuracy can be achieved with a considerable decrease in the total tracking area of the network.
Design of the service-based architectural framework for the South African National Park System
(2004) Khumalo, Themba Cyril; Adigun, M.O.
A typical response to the challenge of rendering competitive services to the customers of the South African National Parks led to the development of architectural mechanisms for providing services by taking advantage of the dynamic web protocol standards and frameworks. This was done in three steps: (1). The evaluation of existing IT-level support in providing nature conservation information and marketing of National Park services; (2). Investigation of mobile commerce services to create customer values that promote customer loyalty based on enterprise values; and (3). The development of a distributed service-based architectural framework for the South African National Park system based on service-oriented architectural model. The building blocks (publishing, registration, personalization of services, etc) of the architecture serve as the basis of designing system services for disseminating nature conservation information and marketing the services of the national parks. A prototype of the information service system was used to prove the usability of the architectural framework developed. The architecture proposed in this work is a guide and should provide the basis for an ICT infrastructure that responds to the quest for modernising the national parks information system.
Enterprise component architecture for mobile commerce services
(2004) Kunene, Hlengiwe Pinky.; Adigun, M.O.
This research focuses on creating a component based repository architecture for mobile commerce services called (e-TOCOR) with the emphasis on component storage and retrieval. To realize this framework three tasks were carried out namely (i) a model for engineering component based m-commerce service was defined using existing models (ii) the Universal Description Discovery and Integration (UDDI) was used to model a component repository (hi) the mobile travel reservation application prototype was developed to demonstrate the proposed model. The results obtained were threefold (i) by evaluating the existing component based architectures, the study showed that m-commerce services are not the same as e-commerce services, the Information Requirement Elicitation (IRE) was adopted as a mechanism for eliciting a request and a service delivery protocol for end-user mobile commerce services and (ii) the prototype was developed to show how enterprise components can be delivered in mobile devices using the IRE protocol. It was also shown that the way existing m-commerce services elicit requests takes much time, the shortest way was to use a text message (iii) the repository framework was created emanating from the home-based reference architecture. In conclusion* the proposed repository could not be compared with the existing repository architecture because it was not implemented, instead the UDDI was used.
Provisioning of secure Multi-Agent Systems (MAS) based on trust contracts
(2004) Biyela, Philile Diligent; Adigun, M.O.
This research work focuses on the development of an Interaction Security Contract (ISC) mechanism that enhances the capability of agents to protect themselves against compromised entities in Multi-Agent system. To realize the framework, two tasks were carried out namely ® a simulation of an agent based system was created and trust accuracy levels based on the proposed mechanism were evaluated to establish the significance of the proposed security scheme; and (if) a security-aware agent-oriented shopping mall system was designed and implemented to demonstrate the proposed contract based security model. The results obtained are threefold; (i) This study was able to establish that trust models based on reputation do not obtain accurate trust values for agents to make correct trust decisions in agent systems; (ii) it was established that it was even riskier to form collaborations without any trust model in place as agents collaborated with agents whose security status was not known to them. This clearly shows that agents were at risk of being compromised; and (iii) it was established that if agents publicize their security properties truthfully, trust in the overall agent systems will be gready improved as agents themselves will only be collaborating with agents who meet their specified level of trust In conclusion the study advocates the employment of the proposed Interaction Security Contract (ISC) mechanism, since it was demonstrated how knowledgeable trust relationships can be formed and improved in agent based systems.
A requirements specification model for a product line
(2005) Kabanda, Salah Kemilembe
This research work focuses on developing a new requirement engineering model (RSPL) based on a Model Driven Architecture and Web-tier application framework, to support automatic and interactive requirements generation and specification when creating families of systems. In realizing the model, two goals were targeted namely (i) to construct a requirement engineering model that support automatic transformation of domain features into actor-specific requirements; and (ii) to design and implement an interactive web based requirement engineering tool that demonstrates the requirements generation and specification process for a product line. The result obtained is twofold: (i) a model driven architecture for rapid requirements generation and specification for a product line that reduces costs and development time; (ii) tool implementation based on a web tier application framework that supports different client and actor types. In conclusion, the study is a contribution to a recently advocated idea that requirements generation and specification for product line development could be model-driven. The result shows that the idea is promising with respect to requirement reuse and improving communication barriers among members of a system development team.
Resource management in the future internet
(2005) Mtshali, Mbalizethu Zandile Ntombifuthi
The fast Internet growth has led to high demands for QoS provisioning mechanism to be implemented in the future Internet This dissertation presents the development and implementation of a resource reservation scheme to allocate and manage resources in the future internet. Resource reservation as one of QoS techniques used for QoS provisioning in the network is crucial for many newly developed user applications to flourish and prosper. The Scalable Efficient Reservation Protocol (SERP) is proposed to efficiently allocate and manage resources in the future Internet. The proposed scheme, Scalable Efficient Reservation Protocol (SERP), takes into consideration the fact that the future Internet will accommodate both real-time and non-real-time applications. Relevant design goals and principles were identified and these led to a formulation of a corresponding model. A simulation of the model was conducted to evaluate the performance of the developed scheme. The experimental results do show that SERP reduces the router load with less states information to be maintained. Use of aggregate-based reservation further reduces states information. Consequently SERP improves scalability. The results further demonstrate that SERP has high request acceptance probability and more connections are established for end-to-end reservation.
Mobility management in wireless IP networks
(2005) Ngwane, Mathalenta Jane Rosa; Adigun, M.O.; Ojong, G.E.
Mobility management plays a significant role in the current and the future Internet in delivering effective services to the mobile users on the move. This research proposes a mobility management architectural framework that will incorporate wireless mobility architecture and call admission control. To achieve this we carried out three tasks, (i) the architectural basis for the mobility management strategy for real-time data communication was established, (ii) a suitable call admission criteria and a corresponding call admission control algorithm for the strategy was defined, (iii) a simulated performance evaluation of the strategy was developed. The results obtained are as follows: (i) this study resulted in a strategy that prioritizes calls of the guaranteed services class over calls of other classes, (ii) the higher priority of handoff calls over new calls has been achieved; and (iii) the proposed Call Admission Control (CAC) algorithm shows to have an acceptable performance with an average acceptance rate of calls of up to 83%. In conclusion this study resulted in a CAC algorithm that prioritizes real-time traffic over non real-time traffic as well as handoff calls over new calls.
Interface strategy pattern for context-aware service based environment
(2005) Zuma, Siphesihle Mandlakhe
This research work focuses on the development of the Context-Aware Component Interfacing Pattern, CACIP, aimed at contributing to current developments in distributed communication and collaboration in mobile wireless environment. CACIP derives from the component bus and the component glue patterns as standard solutions to robust producerconsumer interaction in a mobile wireless environment, Communication between the participating components is managed by a common data structure referred to as rhe information bus. To showcase usefulness of the pattern, a prototype of an example service was implemented which was subsequently used as a testbed when evaluating the performance of the proposed scheme. We evaluated performance of rhe model in terms of mcssage throughput, round-rrip time and we finally measured rhe significance of context-awareness to the model. For message throughput, the model displayed a normal behaviour of any distributed system since as the number of components participating in communication activiries were increased, rhe overall message rhroughput dropped. When measuring the moders round-trip rime, i.e. the time taken to send a request message and receive back a response, we observed that increasing the number of messages exchanged among the participating components caused the mode!'s round-trip time to increase. Again, rhis was an expected resulr. An experiment aimed at measuring the significance of context-awareness to rhe model was also conducted. The obtained results showed rhat when contexr data was not utilised. the number of inaccurate responses was higher compared IQ that of accurare ones. On Loe other hand, incorporaring context-awareness produced rhe opposire of rhe first resulr, rhus proving that comeXI-awareness is a useful fearure rhar mobile apphcations must adopt to improve [heir quality of service.
Handoff management in the mobile wireless Internet
(University of Zululand, 2006) Cwele, Bheki Agrippa
The large number ofnodes in wireless networks and the fact that mobility is an important feature of a wireless system has made it a great challenge to sustain ongoing connections during handoff events. Real-time applications (e.g., Voice-Over-IP (VoIP), Video-Ondemand, and streaming) require little or no service disruptions during handoff because of their nature of being mostly delay-sensitive applications. The network must, therefore, offer minimal handoff delays, minimal latency, low call dropping probability; and should also fully be supportive ofspeed variations in mobile nodes during network layer handoff. In our research we developed and simulated a fast network layer (subnet-to-subnet) handoff management scheme, known as Sector Aware handoff(SAH) strategy. SAH uses cell sectors marked into regions. The sectors have non-handoff region, preparation region and handoff execution region. The SAH scheme uses mobile node's real-time mobility parameters over cell sector regions, to provide accurate and early predictions of Network Layer (L3) handovers - thus initiating L3 handoff preparation appropriately early. SAH is proposed to support primarily, real-time applications' handoff requirements and also fast moving mobile nodes (that is, supports speed variation in mobile nodes). As a means of evaluating the performance of the proposed scheme, a simulation was conducted. The experimental simulation results show that SAH performs better in terms of providing low call dropping probability and maximum system utilisation, when compared to the Fast Hierarchical Mobile IPv6 (FHMIPv6) handoffprotocol.
Guaranteed service management in mobile wireless Internet
(2006) Nyandeni, Thulani Clement.; Adigun, M.O.; Emuoyibofarhe, J.O.
The ubiquity of the mobile wireless Internet has led to the increased use of reatime applications. Real-time applications are network resources hungry and propagation delay sensitive. These applications need service guarantees if they are to be acommodated on the mobile wireless Internet. It is, therefore, necessary to maintain the required QoS of these applications in the presence of user mobility with the use of resource reservation mechanism. This dissertation presents development of a resource management mechanism to allocate and guarantee the availability of network resources in the mobile wireless Internet. A model tagged "Dynamic Mobile Resource Reservation Protocol (DMRSVP)" was proposed. DMRSVP comprises of the following: resource resenaiion, bandwidth allocation, mobility management, and next cell prediction components. The sectored cell approach next cell prediction scheme was modified by adding a resource reservation module. The resource reservation of DMRSVP is made up of a dynamic comparison algorithm for reserving resources in an exactly cell in which a mobile node will visit with open connection. The proposed model utilises HMlPv6 for managing mobility within a domain and the cooperation of hierarchies was adopted to manage mobilyt between domains. An algorithm tagged "connection swapping" was developed for DMRSVP to reduce both call blocking and call dropping probabilities. The simulation of DMRSVP was carried out using an object oriented programming environment, VB.net. The performance of DMRSVP was compared with the existing schemes (MRSVP and HMRSVP) through simulation. The network load, delay, call blocking probability (CBP) and call dropping probability (CDP) were used as the simulation parameters. The results of the simulalons show that bandwidth utilisation of the three schemes between 0 and 2 seconds (sec) were the same, but after 10 sec, MRSVP utilised 160Kbps of bandwidth, while HMRSVP and DMRSVP utilised 138Kbps and 102Kbps of bandwidth respectively. This verifies thaDMRSVP does not over utilise bandwidth. The proposed model (DMRSVP) blocked 5 reservation requests, while MRS VP and HMRSVP blocked 36 and 24 reservation requests respectively within 10 seconds. The signalling overload of the three protocols shows that MRVP sent 225; HMRSVP sent 151, while the proposed DMRSVP sent 25 control messages over the period of 10 minutes. This implies that the proposed model does not overload the network with a control messages; this allows for more user data packets to be transmitted. The round trip delay of reservation set up of MRSVP was 45 sec; HMRSVP was 31 sec; while it took only 19 sec for DMRSVP to set up the reservation. CBP and CDP of the proposed model were measured. The simulation results show that DMRSVP has CBP of .02 and CDP of 0.05. These are lower values compared to what is obtained in the literature for other schemes. Based on the results of the simulation, it was concluded that DMRSVP is more scalable and efficient than the classical MRSVP and HMRSVP. This isdue to the fact that DMRSVP takes into account mobility management but MRSVP does not. Again DMRSVP was integrated with the next prediction algorithm to predict the cell(s) in which a mobile node might visit. This makes DMRSVP not to over reserve the limited network resource.
Modelling trust management in agent-to-agent security scenario
(2006) Nkosi, Mzomuhle Thuthuka; Adigun, M.O.
Over the previous years, researchers have given much attention to explore the effective use of distributed systems to search, retrieve, and share information. Mobile agent technology has had a great impact in providing solutions in different dimensions of distributed systems. This research work presents the development and implementation of trust management model that ensures agents communicating in a Context-Aware Component Interfacing Partem (CACIP) based Service Oriented Distributed Environment. The Agent-to-Agent Trust Management model is proposed to ensure security of agents that want to access services from the CACIP. The developed model uses a reputation-based approach to evaluate tmstworthiness of each agent. Each agent must be authenticated and authorized in order to get an access to services in CACIP. The trust management model ensures trustworthiness of each agent participating in a CACIP based environment. Mobile agent security requirements were considered in developing the model. The developed model was implemented to demonstrate how requesting agents are authenticated and authorized before accessing services. Simulation of the model was also conducted to assess the performance to the model. The results show the efficiency of the model when agents simultaneously make requests over a given amount of time. The performance of the model shows scalability when number of requests increases. The use of both foreign and local reputation to evaluate trust of agents guarantees that all agents that access services, are trustworthy.
Modelling personalizations in the design of mobile publish/subscribe architectural framework for South African National Parks
(2006) Linda, Mcebo Brendon; Adigun, M.O.
The dissertation presents the adoption of publish/subscribe pattern as the existing communication paradigm into the South African National Parks (SANPARKS) organization. It focuses on restructuring the existing SANPARKS system into a publish/subscribe information provider that would enable clients to receive personalized information services delivered via sms, mms and email. A publish/subscribe architectural framework, which supports both mobile and desktop users with the following features was developed: a national park system restructured into an information provider; a portal interface for information consumers; a mobile interface support achieved by personalization and a highly rated usability index. This research was conducted by firstly, surveying the theory which consisted of the literature review. Secondly, the formulative aspect which consisted of model building, the proof of concept such as prototype and the usability testing of the prototype. The results of this research testify that the newly restructured SANPARKS meet the standard of the information provider entity. It also provides a portal interface for information consumers enabled by personalization. From the results obtained in this research the adopted publish/subscribe pattern solves the problem of restructuring the SANPARKS system into an information provider that enables clients to receive personalized information services delivered via SMS. MMS and email.
The effect of topology control for wireless multi-hop networks
(2007) Mudali, Pragasen; Adigun, M.O.
Wireless multi-hop networks are not restricted to rural development efforts. They have found uses in the military, industry, as well as in urban areas. The focus of this study is on stationary wireless multi-hop networks whose primary purpose is the provisioning of Internet access using low cost, resource-constrained network nodes. Topology control algorithms have not yet catered for low cost, resource-constrained network nodes resulting in a need for algorithms that do cater for these types of wireless multi-hop network nodes. An algorithm entitled "Token-based Topology Control (TbTC)" was proposed. TbTC comprises three components, namely: transmit power and selection, network connectivity and next node selection. TbTC differs significantly in its treatment of the synchronisation required for a topology control algorithm to work effectively by employing a token to control the execution of the algorithm. The use of the token also ensures that all the network nodes eventually execute the topology control algorithm through a process called neighbour control embedded within the next node selection component. The proposed topology control algorithm, TbTC was simulated using ns-2 and the performances of a 30-node network before and after the algorithm was utilised, were compared. The Packet Delivery Ratio, Delay. Routing Protocol Overhead and Power Consumption were used as the simulation parameters. The neighbour control process was found to significantly reduce the number of hops taken by the token to visit each network node at least once. It was found that this process shortened the token traversal by 37.5%. Based on the results of its simulation, TbTC proves the positive benefits that can be accrued to the use of tokens in topology control as well as highlighting the negative benefits of the creation of uni-directional links in wireless multi-hop networks that utilise the IEEE 802.11 standard.
Design of partnership-centred information repository
(2007) Okharedia, Joseph Ayangbeso; Adigun, M.O.
Organizations require the development of Information Repository to improve their products and services. With the rising interest in the semantic web, research efforts have been geared towards the integration of distributed information sources into an interoperable knowledge-base environment. The objective of this research was to design an information repository to facilitate the integration of operational data from various sources into a single and consistent knowledge base that supports analysis and decision making within and across different organizations. The research activities that resulted in the proposed information repository consisted of two major steps: (i) the design of an information repository aimed at promoting the usage of e-commerce through the Internet amongst the Small Medium Enterprises and the larger organizations and (ii) the implementation of repository architecture to facilitate knowledge sharing in response to queries on distributed metadata and easily interoperable sources merged into one semantic entity. The SMEs and the larger organizations are focused as partners in the design of this information repository. The implication of focusing on the SMEs and the larger organizations is in line with the South African Government efforts to promote Black businesses, because the former is considered to be one of the most viable sectors with economic potential growth. The functionality of the information repository is partitioned into a set of services which include: (i) registration of SMEs' profile, (ii) service advertisement and (Hi) service delivery. These services provided the mechanism for storing, retrieving and updating of information. Richards Bay Minerals, an organization in the mining industry that has outsourced part of her core business to junior partner was used as a case study to illustrate the model. XML, the Extensible Mark-up Language provided the descriptive language for the exchange of information between the different organizations via the web. The Java programming language was adopted for implementation. A performance evaluation of the information repository based on a set of parameters like usability, scalability, functionality and collaboration was carried out, with the aim of determining areas where improvements could be made in providing solution to the design of information repository for the web environment.
User preference mining for context-aware M-services applications
(2007) Jembere, Edgar; Adigun, M.O.; Xulu, S.S.
Challenges to the field of Human Computer Interaction (HCI) arising from the emergence of mobile computing can be solved by tailoring the access and use of the mobile services to user preferences. User preferences are traditionally assumed to be static, but due to the dynamic nature of the mobile computing environment, this assumption no longer holds. In an m-Services environment user preferences are not only transient, but they also vary with the changes in context. Furthermore, the assumed preference models do not give an intuitive interpretation of a preference and lack user expressiveness. To address these issues, this research work defines a user preference model for a context-aware m-services environment, based on an intuitive quantitative preference measure and a strict partial order preference representation. We present some user preference mining algorithms and a framework for context-based user preferences mining in an m-Services environment. The developed user preference modelling and mining framework was prototyped and evaluated it terms of its quality and effectiveness. The user session data for the evaluation of the framework was generated using MATLAB 7.1's Generalised Pareto Probability Density Function (gppdf) with shape, scale and threshold parameters of 1.25,1, and 0 respectively. The framework was found to be relatively promising in terms of its effectiveness. The user preference mining framework was also found to relatively scale well with increases in the volumes of data.
Gauranteed real-time delivery of context-aware messages in publish/subscribe system
(2007) Shabangu, Petrus Sipho; Adigun, M.O.
Publish/subscribe communication paradigm is becoming popular with its decoupling factor, and the filtering of the message stream during the process of dissemination. In a publish/subscribe communication model, a subscriber is decoupled from the publisher, in the sense that a publisher and a subscriber are physically separated from each other. This dissertation reports an ongoing attempt at guaranteeing real time delivery of messages for publish/subscribe systems in a mobile environment where subscribers continuously change location. It focuses on ensuring that messages are delivered in time and space, do not permit stale messages to be delivered but allows subscribers to select priorities based on their preferences. This research was conducted by first, surveying the theory which gave rise to the theoretical framework used for literature review. Second, the research formulation activity consisted of model building, proving of the crafted model by using a prototyped scenario in which the proposed message delivery architecture was demonstrated and finally, a simulation that tested the reliability of the message delivery model. The results obtained from this research testified that the emanating message delivery architecture ensured that messages are guaranteed to be delivered to registered subscribers based on subscriber preferences.
Location-based mobile E-commerce for tourist services
(2007) Mzila, Phiwa Douglas; Xulu, S.S.; Adigun, M.O.
There are many instances where tourists become strangers in foreign countries and have no way of discovering available services in place they visit. These services, which are known as location based services, are normally available to only subscribers of those services. For mobile users who are not subscribers this may be unfair, especially if a mobile user is new in that environment, as he or she remains unaware of services that are available. This research proposes a dynamic service discovery mechanism whereby every mobile user is capable of service discovery irrespective of being a subscriber or not. We developed a model for location-based service discovery in a mobile environment. With this model we were able to show how components of a location based application server in mobile commerce application could be modelled in order for them to perform their task, and how services could be advertised in the vicinity of the network coverage area by the LBAS so that the user who is entering that area would be able to access services that are available to him/her We then implement a prototype of the model to showcase the usefulness of the proposed model. Our last objective was to evaluate performance of the prototype system. We conducted usability testing where we asked different people, from different faculties, to take a tour of the system. This group was then interviewed to evaluate the user-friendliness and the easiness to learn testing of the system. In both aspects, the users were satisfied with system.
A dynamic and adaptable system for service interaction in mobile grid
(2007) Moradeyo, Otebolaku Abayomi; Adigun, M.O.
Mobile and pervasive computing with its peculiar feature of providing services at anywhere anytime basis has been at the centre of major computing researches in recent times. The device resource poverty and network instability have been reasons behind unsuccessful use of handheld technology for service request and delivery. However, interaction of these mobile service components can be adapted to further improve on the quality of service experienced by service consumers. Content, user interfaces and other adaptation mechanisms have been explored, but these have not provided needed service qualities. However, one of the challenges of designing an adaptable system is on making adaptation decisions. This dissertation, therefore, presents a dynamic and adaptable system for service interaction. A context-aware utility-based adaptation model that uses service reconfiguration pattern to effect adaptation based on contexts was developed. It was assumed that developers of mobile services design services with variants that can be selected at runtime to fit the prevailing context situation of the environment. All variants differ in required context utilities. The service variants selection decision is based on a heuristic algorithm developed for this purpose. A prototype of the model was built to validate the concept. Experiments were then conducted to evaluate the proposed model purposely to measure the interaction adaptation quality, the overall response time with or without adaptation, and the effect of service consumer preference for a given service variant on adaptation process. Results from the experiments showed that though the adaptation process comes with additional overheads in terms of variation in response time, the adaptation of service interaction is beneficial. It was observed that the overall response time increased initially as the number of service variants increased which was due to overheads by the adaptation process. However, as the number of variants increased, the response time began to fall sharply and then became steady. This proved that adaptation can actually help reduce service response time. We also found that adaptation quality degraded with increased number of service variants. The lesson learnt was that adaptation can help reduce overall response time and can improve service quality perceived by the service consumers.
Dynamic multi-target user interface design for grid based M-Services
(2008) Ipadeola, Abayomi Olayeni; Xulu, S.S.; Adigun, M.O.
Device heterogeneity and the diversity in user preferences are challenges, which must be addressed in order to achieve effective information communication in a mobile computing environment. Many model based approaches for the generation of user interfaces have been proposed. These approaches were aimed at achieving effective information communication in a mobile computing environment. Unfortunately, current model based approaches do not support user participation in interface adaptation. Hence, user interfaces are adapted for and on behalf of users, leaving users with inappropriate interfaces. There is therefore a need for an approach that allows direct user participation in interface adaptation. Such an approach must be responsive to changes in user needs and preferences. This work therefore proposes a Polymorphic Logical Description (PLD) method for automatic generation of multi-target user interfaces. The PLD method consists of user preference aware models, which are created at design time. Interface artifacts are considered in the PLD as methods with polymorphic attributes in a bid to address the diversities experienced in a mobile computing environment. An interface artifact is dynamically selected for inclusion on an interface based on the context information of a requesting user and intrinsic characteristics of a device. Our approach is enabled by a support toolkit, named Custom-MADE (CoMADE) for automatic interface generation. The evaluation result of CoMADE shows that it achieves a high degree of user participation during interface generation, flexibility, dynamism, ease of application extensibility and user centeredness interfaces.
Design of an IP address Auto-Configuration scheme for Wireless Multi-Hop Networks
(2008) Mutanga, Murimo Bethel; Xulu, S.S.; Adigun, M.O.
The importance of wireless ad-hoc networks (eg wireless mesh networks) in community and commercial connectivity cannot be underestimated in view of the benefits associated with such networks. An ad-hoc network must assemble itself from any devices that happen to be nearby, and adapt as devices move in and out of wireless range. High levels of self organization will minimize the need for manual configuration. In essence, self-organization provides an out-of-the-box functionality such that very little technical expertise is required to setup a network. However, efficiently providing unique IP addresses in ad-hoc networks is still an open research question. The goal of this study, on wireless multi-hop networks, was to develop algorithms for IP address auto-configuration. These algorithms should address the following among other problems: Achieving high levels of address uniqueness without compromising on latency and communication overhead. To achieve the overall goal of this research we proposed changes to the traditional DAD procedure, the Wise-DAD protocol was proposed. We introduced state information maintenance, which is passively collected and synchronized. Passively collecting state information reduced the number of DAD trials thereby reducing latency and communication overhead. Simulations were done in NS-2 to test the performance of the proposed protocol. A comparative analysis was then conducted. Wise-DAD was compared with the Strong-DAD protocol. Experiments on the effect of network size, node density and node arrival rate on communication overhead, address uniqueness and latency were conducted. Results from the simulation experiments show that Wise-DAD outperforms Strong-DAD in all the three metrics used for performance evaluation. First, Wise-DAD showed better scalability since it performed better than Strong-DAD when network size was increased. Communication overhead in Wise-DAD was generally low whilst the latency was generally uniform. The number of IP address duplicates recorded was reasonably low. Second, Wise- DAD was not affected by node arrival rate on all the three metrics that were recorded. On the other hand, the number of address duplicates in Strong-DAD decreased as the node arrival rate was increased. Interference significantly affected communication overhead recorded in Strong-DAD. Wise-DAD, on the other hand, was not affected by interference. The number of address conflicts in both protocols showed an inverse relationship to interference. However, the number of conflicts for both protocols was significantly different; Wise-DAD recorded much less address conflicts than Strong-DAD.

Browse

Browsing Computer Science by Issue Date

Results Per Page

Sort Options