IP Broadcasting over Unidirectional Satellite Link

IP Packet Loss and Recovery over Unidirectional Satellite Network

     Unidirectional satellite communications becomes much popular because of wide geographic coverage, multicasting features. These features support several one-to-many applications where asymmetric connection is sufficient. At the same time, satellite communications have some disadvantages like inherent noise in the channel or high loss rate because of some weather condition like rain, fog etc. Several multicast transport protocols have been proposed to these applications for UDL where reliability is maintained by a narrow backchannel. Much less works have been devoted to maintain reliability in UDL without backchannel where many receivers are concerned. In this situation, it becomes very much necessary to study about the packet loss patterns in different satellite receivers and how efficiently the packets can be recovered among recipients without contacting to the sender. In this study, a detailed analysis of multicast packet loss patterns is shown with experiments in a real satellite network test bed. An analysis for average number of packet loss by many receivers is also shown those can be recovered by the receivers communicating among themselves without requesting for the lost packets back to the sender. A simple bitmap data structure (fingerprint) named Mixed-PEDbit is designed for the receivers as if they can exchange the packet information (received or lost) with other receivers (peers) with reduced communication. Additionally, a recovery framework is recommended to apply these bitmap data structures.

AI3 Networks

Multicast Push Cashing System

      The World Wide Web (WWW) continues to grow at an exponential rate, leading to the problems of access latency and overloaded servers. The distributed cooperative caching concept is applied throughout the Internet community to reduce bandwidth usage, improve web server availability and improve user access latency. The distributed cooperative caching system is a group of web caches sharing their contents and workloads. However, cache consistency could not be satisfactorily manageable. Some cached objects are not up-to-date, thus they could not serve user requests immediately. Multicast push technology is proposed to improve the quality of web caching system. WWW documents are continuously delivered to all member caches by reliable multicast transmission. Analysis of user access patterns is designed to accurately predict the objects which are going to be referenced in the future. Concerning bandwidth optimization, caches are designed to dynamically subscribe and unsubscribe to the push operation depending on their interests. Experimental results obtained from this study demonstrate that the multicast push caching system improves the performance of the original cache system. Extra bandwidth usage is spent on data distribution to update cached contents. The proposed system can effectively support the scalability of the cooperative caching sytem because the data distribution is based on multicast technology which provides optimal bandwidth usage.

Scheduling Mechanism for Multicast Transmission in WWW Caching System

      A system of distributed cache servers connected together hierarchically has been widely adopted in practice today. At the top of the hierarchy, a large cache server called hub cache retrieves information from the original source or from some other high-speed caches to feed a community of somewhat smaller children, here referred to as rim cache. Normally, the rim cache sends WWW document requests to its parent hub cache. The hub cache retrieves the requested WWW document from its own cache or from the original WWW server and then sends the requested documents to the requesting rim caches. To enhance the cache performance by distributing WWW objects among the member caches, we propose to add a new feature where the hub cache multicasts popular WWW document obtained to all connecting rim caches regularly, so replicated copies are kept on the member rim caches. In other words, the hub cache pushes popular documents to all members while less popular documents are fetched from the hub cache by the rim caches upon demand (pull). Replicated objects are thus kept closer to the clients improving the cache hit rates and reducing the WWW access time at the clients. To multicast web objects continuously over the network, one is risking the fact that multicasting packets could easily flood the network. Thus a scheduling mechanism is required to manage the multicasting process taking into consideration the error control, the bandwidth consumption, traffic congestion control and the scope of the multicast transmission from hub cache to all connecting rim caches. In this thesis, an adaptive sender based approach is proposed and implemented. Experimental results have confirmed that this approach significantly improves the transmission time, consumes low and static bandwidth consumption, as well as has low retransmission ratio.

Reliable Multicast over Unidirectional Satellite Link

Packet Delivery Scheduling for Reliable Multicasting over Unidirectional Link

     Reliable Multicast is encountered in many applications where there is a need to disseminate information to a large user community in a wireless asymmetric communication environment. To utilize the transmission media effectively, it is necessary to have efficient scheduling algorithms that can balance individual and overall performance, and also can scale in terms of client population and transmission bandwidth. This paper considers the problem of multicasting over unidirectional links that did not provide any feedback channel. The way to provide the reliability mechanism is to transmit data consecutively until the end of available time period. This study proposed  Multi-level Queue for the scheduling algorithm and separated each data by the priority. The scheduling used the scheduling pattern to control data transmission of each queue. Finally, the experiments had been conducted to demonstrate the achievable performance of the transmission scheduling. It was found that the performance of the scheduling depends on many factors such as scheduling pattern, available network bandwidth, etc. The main factor is configuring of the transmission parameters to match with the system bandwidth. The inaccurate configuration can lead to the congestion problem and diminish the overall performance.

Flow Control for Reliable Multicast over Unidirectional Links

       This study designs and implements flow control mechanism for reliable multicast over unidirectional link. The current reliable multicast protocols are not designed for asymmetric bandwidth network, such as unidirectional link, which causes the congestion at the return path. The study therefore proposes the RMU (Reliable Multicast for Unidirectional Link), which is based on the sender-based approach, to provide the reliable bulk data transfer in the multicast environment over unidirectional links. The sender sends monitoring packets in order to monitor the network or receiver status and adjusts the transmission rates using the receivers reception results. To improve the bandwidth consumption, the RMU uses the selective repeat approach, which retransmits only the lost packets. Furthermore, the feedback implosion problem can be solved by the backoff and polling mechanism.

Scalability Analysis Of Reliable Multicast Protocol For A Unidirectional Satellite Link

     The development of a category of applications that simultaneously connect different users scattered over different locations has fostered a new kind of transport protocols: reliable multicast protocol. Satellite links have been viewed as the potential media for broadcast and multicast communication and are utilized to transmit application data to multiple users simultaneously over very large geographical area. Reliable Multicast protocol for Unidirectional Satellite link (RMUS) is a transport protocol which ensures reliable data transmission to many users and is designed for utilizing unidirectional satellite links. Scalability of reliable multicast protocol is one important criterion that needs to be studied to foresee the performance of the protocol under the pressure of a large number of users. The thesis is aimed at studying the scalability of RMUS. The protocol has been modeled using simulation and the performance data is analyzed to apprehend an insight of the RMUS protocol. Simulation data shows that the implosion phenomenon leads to the increase in the transmission delay and the decline of throughput. Other performance figures such as packet latency and overhead are also impacted. The impact of a large population of users on RMUS with various timeout intervals is investigated where it was found that in order to accommodate larger groups of receivers the timeout setting should be enlarged. A method to compute the minimum backoff interval that is used in setting RMUS timeout is proposed. To keep the protocol intact the method figures out the backoff interval for a multicast network including the scenarios chosen for running RMUS. The injected flow of acknowledgement packets through the Internet is simulated to investigate the impact of backoff interval and traffic on the drop of ACK packets. The minimum backoff is large but it still needs to be increase in the presence of traffic. The loss of ACK packets on the return path to the server leads to the false alarm of the congestion on the satellite link. The adjustment of backoff to increase ACK received by the server can lead to more realistic feedback on the network condition.

Reliable Content Delivery over Unidirectional Link using IP Datagrams

       Sending and receiving data over Unidirectional Link provide an alternative choice of delivery. With this architecture, we can send data to any places where there exists a satellite or radio signal. Since unidirectional link do not have any feedback from receiver, there is no mechanism to ensure reliability of data delivery. This thesis applies a more reliable delivery approach in unidirectional link by using duplication technique. Receiver is guaranteed to reassembly all data packets to get the original data set even if data packets had boisterous arrived to receivers. Experiments were setup in different environment to measure receiving capability and efficiency in restoring original data set. Experiment results in this study have exposed that; data delivery over unidirectional link could be completely received within an appropriate time if packet loss rate is low. It was found that bandwidth and speed of CPU is significant impact to the performance of the application. In case that bandwidth was not enough, packet lost will immediately occurred because some data packets have to be discarded. Speed of CPU was another reason of data lost; if receiver spent too much time for reassembling data packet, then receiver would miss the following data packet.

Wireless Unidirectional Web

      With a tremendous growth of interest to access the Internet via mobile devices, Wireless Application Protocol (WAP), has been introduced as a bridge to connect between mobile Network and the Internet world. With the current WAP technology, a WAP gateway is employed to translate data between regular mobile packet to the Internet protocol. If millions of users of the mobile network are moving to the Internet technology, WAP gateway can easily become overloaded. As a result, this thesis proposes to deploy the concept of unidirectional traffic for mobile web access with multiple WAP gateways. With the new system, the return packets from the web server do not have to be channelled through the same WAP gateway. Improvement on average response time per each client s request was observed in the new system. The number of gateways has been shown to effect the average response time and the new system dramatically improves the overall performance in case of large number of hops.
 

Gateway Selection Based On Gateway Load In Ad Hoc Infrastructured Environment

       Mobile Ad Hoc Networks are standalone networks that are combined with the fixed infrastructure network to create an Ad Hoc/Infrastructured environment. This environment consists of multiple gateways that act as entry/exit points for the traffic flowing between the MANET and fixed infrastructure network. As, there are multiple gateways, therefore, there is an option to select one of the gateway from all the available gateways based on some criterion. This research proposes a solution other than hop count to select multiple gateways in such an environment. The proposed solution considers the load on the gateway as the criterion for selection in which the gateway with lighter load gets to be selected. Furthermore, it is compared with the hop count variation based on the mobility in the Mobile Ad Hoc Network. The results show that the proposed solution performs better as compared to the hop count variation in terms of average end-to-end delay but suffers in case of the number of packets delivered to the destination as the protocol overhead for the proposed scheme is higher.

A Practical Approach Towards Server-Less Group Member Authentication Protocol For A Mobile Ad Hoc Network

     A wireless mobile ad hoc network (MANET) is a collection of wireless mobile nodes dynamically forming a temporary network, without the use of fixed infrastructure. Zero setup cost and dynamic topology of MANET makes it an attractive choice for group-oriented applications. But many limitations such as lack of infrastructure, less bandwidth and memory, wireless link, noise, etc. in mobile ad hoc networks impede the usual deployment of traditional security infrastructure. Moreover, according to the characteristics of transiently associated network, security incident in MANET mainly binds to group communication. To address these problems, in this thesis, we have designed a protocol for secure group communication in MANET. Our protocol feasibly works in a server-less environment and provides a distributed lightweight authentication mechanism for instantaneous deployment. Its core technique consists of zero knowledge proof and threshold cryptography.

An Implementation of Wireless Internet Sharing Plug-in for OLSR

     In an office or campus environment, a very common method of connecting computers together is the IEEE 802.11 b/g family, with base stations, and hosts connect in managed mode. This is an infrastructured approach, with certain limitations. Tests performed in the testbed show that the coverage area of the base station is limited, and certain areas of the test bed are not covered by the base station. Depending on the location the actual bandwidth varies a great deal and the diversity of the delay will be significant. An alternative to the infrastructured approach is the ad-hoc approach, where the Mobile ad-hoc network (MANET) is a widely used version of the ad-hoc family. MANETs are usually either not connected to the Internet, or connected by a wired service. To increase the coverage area of the managed mode base station, this thesis merges the two styles of wireless networks, in order to draw on the strengths of both styles. The wireless Internet connection of the managed mode network, and the packet forwarding of the MANET combined in one host . A laptop with two wireless network cards is running OLSRd with the new Wireless Internet Sharing Plug-in (WISP). One of the network cards is connected to the base station, in the managed mode, the other cards is connected to the MANET in the ad-hoc mode. The WISP router can then share its connection to the Internet with the rest of the MANET. The MANET includes packet forwarding, so that the Internet can be forwarded in multiple hops in the MANET. This thesis tested three different modes, the Managed mode, the WISP ad-hoc mode, and a pure OLSRd wired sharing ad-hoc mode. Running three modes of tests enables the thesis to compare the results of the infrastructured managed mode network to the MANET OLSR WISP implementation, as well as a pure OLSR MANET with a host connected to the internet through the LAN. These tests show that the merging of the MANET and the Infrastructured mode can increase the coverage area by more than 100%. The bandwidth of this mode is lower than the managed mode in most locations, but the bandwidth is stable over the whole network, including sites that were outside the coverage are of the managed mode. The delay of the WISP network is slightly higher than the managed mode, but is very stable and shows no locations with very high delay as can be shown in the Managed mode.

Fast Mobile IPv6 for Real-Time Streaming Media

      Mobile IPv6 provides IP layer mobility management for IPv6 networks. Despite many advantages it offers in comparison to Mobile IPv4, handover management still remains an issue for Mobile IPv6. For real-time connections, the handover latency and signaling overhead incurred by the protocol may become significant. This research presents a new method named  Cross-over MAP based Hierarchical Mobile IPv6 (XMAP-HMIPv6) which inherits the advantage of Hierarchical Mobile IPv6 of reducing signaling load for inter-domain mobility. We incorporate an analytical model for the performance analysis of this new scheme along with some existing handover management proposals. Based on this analytical model, we formulate handover latency and signaling overhead costs as evaluation metrics. Finally, we present analytical results in different environments and compare our protocol with the other existing proposals. The analytical model results show that, our scheme performs almost similar to [F+H]MIPv6 in terms of reducing handover latency and outperforms both HMIPv6 and MIPv6.

 

Rural Connectivity by IP Broadcasting over Unidirectional Link

     Satellites enable people in remote areas to receive information and be connected to the world. Due to high cost, the terrestorial communication infrastructures on these islands as well as remote areas on the continent have not been well provided. Satellites are cost-effectively used to broadcast data to areas which lack terrestrial communication infrastructures and even mobile or wireless networks. To deliver contents from the global Internet to the people of these areas, one has to rely solely on the use of unidirectional satellite link to receive information where wire communication infrastructures for sending acknowledgement or feedback to the senders do not exist. With such an environment, there is a challenge in providing a reliable communication service to these areas via the unidirectional satellite link. Our project is to study contraints and requirements needed for designing and implementing an IP-based content delivery protocol over unidirectional sattellite link. On this particular scenario, the protocols must be able to support high-level applications such as bulk file transfer for any media.