This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
The Internet and Internet base technology has become part of our daily lives. Internet base technology is changing the way individuals and organizations are using Internet and Intranet today. Although the present Internet technology was designed for point to point communication (transmission of block of data received to and from source and destination host). The evolving technology demands for real-time distribution of data to a group of recipients located in different parts of the Internet. These applications require simultaneous delivery of data to a set of sites in contrast to the traditional point-to-point delivery used in "unicast" transmission model. The developing technology is known as "multicasting".
"Multicasting" is an efficient way of sending block of data to a set of interest recipients. In contrast to the old "unicasting" and broadcasting delivery method of communication, "multicasting" is more economical and efficient, consumes less processing time and bandwidth, reduced load on sender host, better scalability and does not easily lead to network congestion as the number of client increases.
Organizations are using this "multicasting" application in their operations to send real-time information from headquarters to different branches carrying out their businesses. The "multicasting" applications include data-casting, digital entertainment video and audio distribution, "multi-site" corporate video conferencing, broad distribution of monetary data, grid computing, stock quotes and news bulletins distribution, database replication and software distribution. Internet Protocol "multicast" offers a solution that conserves the network bandwidth by sending the data to multiple beneficiaries without impacting on the sender or the recipients.
Data communication model used in most of our network today are "unicast" and broadcast model of transmission. The "unicast" transmission uses Transmission Control Protocol (TCP) that consume huge amount of network bandwidth and makes packets transmission slow. It waste a lot of processing time of network infrastructures since the router processes each packet going to different destinations one after the other. This multiple processing lead to network congestion especially as the number of clients or recipients increases.
Broadcast model of transmission floods the packets to both the interested and uninterested recipients on the network. The flooding processing lead to inefficient use of network bandwidth, consume routing processing time, and prone the network to congestion easily. This research seek to model a network that will deliver data on to a group of interest recipients using the same best effort as "unicast" model and improve the performance of the network.
Co-operate organization today yearns for real-time applications delivery to multiple site at a reduced cost. The demand for "multisite" co-operate video and audio conferencing is on the high demand. Distance learning and "data-casting" use for training programmes, and educational purposes to deliver the lecture is highly needed. Software update and data replication is extensively used for software patches, and several other evolving applications that demand multiple location real-time delivery. Researches on these solutions are inevitably needed to deliver these solutions, cost effective and improve the performance of the network.
Transmission of heavy data real-time over Internet to a large group of recipients is a huge challenge to the network resource using traditional transmission model. The minimum bandwidth required to transmit heavy data to a large group using point to point transmission model is quite enormous. For instance, a source wants to send data of DVD disk to one million recipients evenly distributed around the globe example IP television. The transmission of the said data will require a large amount of bandwidth (e.g. 100000000bps) which may not be readily available across the network. This will pose a problem of congestion to the network and its infrastructures. In contrast, if the same source could transmit 1000mps of traffic to twenty (20) remote distribution locations, each node can make use of local distribution network bandwidth to distribute the data to fifty (50) thousand subscribers. The central network will requires only (20mbps) to distribute data to twenty (20) distribution points which can be easily implemented across the network and very efficient.
The second motivation is the issue of heterogeneous nature of the Internet. Internet consists of interconnected networks of computers with different network capabilities running different applications. The transmission of packets through different network with routing protocol and how scalable the network will be as the number of clients keeps increasing also motivates this research. The increasing demand by medium and large organisation for efficient and cost effective networks that will consume less bandwidth and reduce load on network infrastructure was another motivation for this research.
GOAL AND OBJECTIVES.
The goal of this research is to design an extensive network that can transmit heavy files like video, audio and database from a source to different locations at the same without additional burden on the network resource. This models simulation will investigate and analyse the results which can be used for future design of "multicasting" network capacity planning. It will also pursue to improve the performance of the network using "OPNET" simulation tool.
In other to achieve the above stated dissertation aim, the following objectives has been set:
1. To study the concept of one-to-many and many-to-many.
2. To study Internet protocol and other layer three routing protocol.
3. To study the different protocol involved in "multicasting".
4. To learn "OPNET" simulation tool.
5. To simulation fixed and wireless network using multimedia and database as inputs data.
6. To compare the result of the network and "multicasting" scenarios.
7. To seek to improve the network.
"Multicasting" delivers data to interested recipients using the same best effort used by "unicast" to send data to a single recipient. The dissertation will focus on the design of a "multicast" network that will improve the quality of the network using shortest path tree and pruning the unused links.
The scope of this research centres on designing different "multicasting" scenarios for fixed and wireless network using different routing protocols for real-time transmission of data over network to sparsely distributed clients on the network. This means that, it will support delivery of data at the same time over heterogeneous network like Internet to set of interested recipients. The "multicasting" network will be using light and high resolution multimedia applications over the Internet. Also, a scenario for database and heavy file transfer over the heterogeneous network will be design to show how these dada are send from a single source to many recipients using file transfer protocol (FTP). The concept of modelling in "OPNET" simulation tool will be used to design and verify the statistics (link utilization, throughput and end-to-end delay) to achieve a cost effective network which is the interesting part of this research.
The limitations in this dissertation are as follows:
1. "IPv4" "multicast" addressing was used in implementation since "IPv6" "multicasting" is not support by the version of "OPNET" used.
2. Pure "PIM-DM" is not supported, due to constraint pose by the present version in "OPNET" used.
3. Security was not considered in this research due to restriction pose by the "OPNET" simulation tool used.
BACKGROUND AND RELATED TOPIC.
In the network, there are certain tasks that must be accomplished before a packet is transmitted from the source node to the destination node. One of it is addressing of the packet - which involves encapsulating of the source and destination address to enable the system know where the packet is going. Another is the transmission of the packet from source to the destination node. That means sending packet from the source to destination. There are many methods of addressing and sending packet over the network. Packets can be differentiated from each other depending on the addresses they carry. The number of destinations host can also be used to determine the mode of transmission used.
There are three methods of transmission of packet in a network namely; "unicast", broadcast and "multicast". "Unicast" is a point-to-point method of transmission. It involves addressing a packet to only one node. Broadcast is a point-to-all method of transmission. That is packet send by a node is received by all the nodes in the network. Finally, "multicasting" is a point to "multipoint" method of transmission. That is packet sent by a node is received by many other nodes.
"Unicast" Transmission Method.
"Unicast" is a one-to-one communication method. The data packets are sends from a source node to a destination node. A typical example is a server and workstation in a network. Each transmission session created involve one server and one recipient per time. Although several nodes may be requesting for the same information at the same time, the server stream or send the data packet to each node one after the other.
Figure 1.0 illustrates the network environment for "unicast" method of transmission. It consists of a server and three end nodes connected in a network. Although the entire nodes requested the same information at the same time from the server, the server sends the data independently using their individual addresses. That is the server encapsulates each packet with end node address and sends it distantly to three nodes. This method of transmission is not scalable especially when the number of the nodes increases. It also consumes network bandwidth as result of redundant packet transmitted on the same path. This transmission method support several standard applications like File Transfer Protocol, Hypertext Transfer Protocol (HTTP), "SSH", "POP3" and simple mail transfer protocol in "IP" network and "Ethernet" network.
Broadcast is a one-to-all communication method. The data packets are sends from a source to all the devices that are connected to the network. The source node sends a copy of the data packet and the routers duplicate and forwards the data packets to the entire connected nodes in the networks. For example, when a network node joins the network first; a broadcast packet is send by Address Resolution Protocol (ARP) to every node on the network for address resolution and to announce it presence in the network. This transmission mechanism floods the entire network with broadcast packet. Unlike "unicast", broadcast transmission method support "Ethernet" network. Broadcast is not scalable on the Internet because it leads to broadcast storm and network congestion. Figure 1.1 illustrates a broadcast transmission in a network environment. It consists of a server and three end nodes connected in a network. The server sends a copy data packet and the router duplicates the data packet and forwards it to all the connected devices. This data packet is received by interested and uninterested end nodes in the network. These unrequested packets received by the end nodes lead to waste of network resource and the flooding mechanism of transmission consume large amount of network bandwidth. Therefore, broadcast method transmission is not appropriate for large networks like Internet.
"Multicasting" is much more different from the "unicast" and broadcast communication methods. It is a one-to-many communication method. The data packets are sends from a source to a group of destination nodes. That is a copy of the data is send from the source host to all the members of the "multicast" group. Internet Protocol "multicasting" communication method also provides many-to-many transmission method. That is group of nodes (minimum one node) can send data packet to a group of recipients. The senders transmit data packet to a "multicast address" that belongs to a particular "multicast group". The data packets are duplicated by the "multicast" routers and forwarded on request to the "multicast" group member that is interested in the data packet. The interested end-nodes that want to receive the data packet have to join the "multicast" group. The dynamics enable data packet to be send only to the interested recipients.
Figure 1.2 illustrates a "multicast" transmission in a network environment. It consists of a source server and three end nodes connected in a network. The source server sends a copy data packet to a "multicast" group address and the router duplicates and forwards the data packet only to the "multicast" group members. It is only the member that joins the group receives the data packet, other nodes who are not member will not receive the data packet as demonstrated in figure 12.
IP Routing Protocol.
Apart from different methods of transmission of data packet discussed above, another important task that must be accomplished before a packet can be transmitted from a source host network to the destination host network is logical addressing. This involves encapsulating the data packet with the source and destination address to enable the system to know where the data packet going.
The logical addressing takes place in the network layer of the "OSI" reference model. These logical addressing and routing data packets are the primary function of devices in this layer e.g., router. The router connects the small networks together to form a huge network that can span cities and the entire world as in the Internet network. This device forwards data from one network to another and allow host on the network to send data packets to one another even if the source host has no idea of the location of the destination host. Router care less about the network host but only cares about the network and best path to each network. To carry out these vital functions, there are set of rules called routing protocol that enables the router to perform these functions.
This set of rules enables the router to find out the best route, and alternative route to every network, and uses this information to pollute the router table (route to every network). For network that are not directly connected to the router, the router must use several ways to discover how to get to the destination network: static (input all network locations into the routing table), default and dynamic routing.
Static routing occurs when the administrator manually add routes to each router's routing table. This routing is very economical, and there is no overhead on the router central processing unit. There is no bandwidth usage among the routers, and it also adds security to the system, because the administrator prevents or allows access to a network. However, it is not scalable in large networks, and demand reconfiguration of the entire routers when there is a change in network topology.
Default routing occurs when data packet are send to remote network that is not in the routing table to the next hop router. Default routing is only used network with one single exit path of the network - stub network.
Dynamic Routing is the used to protocol to locate dynamically network destination and automatically update the routing table. The router learns all the connected networks and other routes that lead to the remote network running the same protocol. The router will sort for the best route to the every remote using the routing tables. The routing protocol will distribute the best route information to every other router in the network using the same protocol, thereby expanding the routing information of the existing network and how it can be reached. This makes dynamic routing to adapt to network topology change, link failure and device failure. This capability makes management, and maintenance of dynamic routing simple compare to default routing, and static routing.
There are two types of routing protocols operating in Internetwork: interior gateway protocol (IGP) and exterior gateway routing protocol (EGP). IGP protocol uses dynamic routing to exchange routing updates within the same domain, while EGP protocols uses dynamic routing mechanism to exchange routing updates between different domains. These protocols operating in "internetwork" use dynamic routing. In dynamic routing, router uses two basic methods to discover the devices, and the route to every device in the network. These methods include distance vector and link state protocols. There is also another protocol, which is the combination of link and distance vector protocol named hybrid protocol. A typical example of hybrid protocol is EIGRP. We will discuss pros and cons of these protocols within a domain (IGP) and between different domains (BGP) in this section. Nevertheless these routing protocols will also be used in our simulation experiment in chapter five and six.
Interior Gateway Routing Protocol.
Interior Gateway Routing is when the routing information is exchange among router within domain. The Interior gateway protocol is categories base on the above routing mechanism namely, distance vector, link state and hybrid routing protocol. The example of these protocols includes Routing Information Protocol (RIP), Open Shortest Path First and Enhance Interior Routing Protocol (EIGRP). These protocols can further be classified as class full and classless routing protocols.
Distance Vector Routing (DVR).
Distance vector uses a distance and direction to determine it best path to remote network. The distance metric is estimated by the number of the hop the packet data will transit. The hop is any time a packet passes router. The "DVR" select it best path based on the route will the least number of hop. The vector is the direction the data packet will transit through the network. The DVR exchange it routing table with only directly connected router. Examples of distance vector are RIP and IGRP. This routing mechanism is also called Bell-Ford algorithm. It is simple to configure with low computational overhead and supportive to troubleshoot.
Link State Routing Protocol.
Link state routing protocol, also called shortest part first works using link state algorithm. The protocol works by updating every router in the domain about other routers and their directly connected networks. In other words, each router has a complete network picture of that domain network and can determine the optimum path to all networks on its own. The protocol enables the router to keep track of information by maintaining three separate tables. One of tables is for directly connecting router, another for the entire network topology, and one for best path (routing table). Instead of exchanging the entire routing table, link state only exchange link updates to each other. Examples of link state protocols are OSPF and IS-IS. In our simulation experiment in chapter 5, all routers in the domain will be configured using OSPF routing protocol. This protocol is scalable in large networks and complex in configuring compared to RIP routing protocol.
Hybrid Routing Protocol.
Hybrid Routing also referred to as balance-routing makes use of the link state and distance vector algorithm in routing in of data packet and in neighbour router relationship. It uses traditional distance vector to update routing information of directly connected network and combine it the cost of reaching them from the perspective of the advertising router. It exchange full routing table with the neighbouring router and send routing update information only when the network topology has changed. Balance-routing protocol converges rapidly, requires less processing time, and memory as compare to link state routing.
Exterior Gateway Protocol.
Exterior Gateway Routing protocol is a protocol that is design to perform routing function between different domains. It can be describe as inter-domain routing protocol that exchange routing information between autonomous systems. An autonomous system (domain) is a network or group of networks operating the same routing protocol and control from a common administration. However, Internet network consists of several autonomous systems, and EGRP is the routing protocol used to route traffic between and through different the Internet Service Provider networks. Examples of EGRP are Border Gateway Protocol (BGP). BGP use path vector mechanism in routing. Path vector routing mechanism enable the node speaker in each autonomous system to create it routing table and exchange withe neighbour speaker node in another autonomous system.
However, before we discuss the interior and the exterior gateway routing protocol, let look at the three types of routing which includes; static, default and dynamic routing.
Static routing occurs when the administrator manually add routes to each router's routing table. This routing is quite economical, and there is no overhead on the router central processing unit. There is no bandwidth usage among the routers and add security to the system because the administrator prevent or allow access to a network. However, it is not scalable in large networks, and demand reconfiguration of the entire routers when there is a change in network topology.