You are here: UK Essays » Essays » Information Technology » Broadband Networks Data

Cookie Information

Privacy Information

Broadband Networks Data

1 - Introduction:

Broadband Networks are networks which carrying multi services Voice, Video and Data. We traditionally have two types of networks pure data networks which are totally packet switching networks and voice networks, telecom networks was traditionally fully voice oriented, but today's networks are broadband networks which carrying voice video and data. Broadband networks are multi media networks which are multi point and multi and multi rate as well. Traditional telephony networks were use to use voice medium only. They were just able to give point to point connection, normally connect only two phones per call and uses fixed bit rate circuit switching.

In modern communication services networks are multi services. Multi media networks may communicate audio, video and data users use these services on daily bases. Users are use to see online movies and there are many full motion video services. "Internet-protocol-based backbone networks are gradually moving towards a network model consisting of high-speed routers that are flexibly interconnected by light paths set up by an optical transport network (OTN) consisting of WDM links and optical cross-connects. Recovery mechanisms at both network layers will be crucial to reach the high availability requirements of critical services. In such a model, the generalized multi-protocol label switching (GMPLS) protocol suite can provide a distributed control plane that will be used to deliver rapid and dynamic circuit provisioning of end-to-end optical light paths."[1]

In both data and mobile networks we have already achieved acceptable data rate. We don't feel any delay to transfer a data file and a simple phone call is very cheap and without any noise. So we have already done a lot of works in both fields. In modern communication networks where we are expecting video calls, live streaming and many more broadband services we still need to improve the networks for QOS. Traffic types in Broadband networks are classified into three characteristic bandwidth, Latency and cell delay variation.[2]

1.1- What is a Router?

Router is device of some time only software in any computer which is responsible to make communication between two different networks. Without router two different networks cannot communicate with each other. Router make it possible it has many other responsibilities like to determine the best path, take a correct routing decision, keep update his database or routing table about the status of the network. In today's network router is key part. The position of router in networks is Gateway; Gateway is points where two are more networks are connected. Router use to connects two LANs (Local Area Networks), LAN and Internet, it connects to WANs or make the communication possible between LAN and WAN.[3]

The point why I need a router is - "For most home users, they may want to set up a LAN (Local Area Network) or WLAN (Wireless LAN) and connect all computers to the internet without having to pay a full broadband subscription service to their ISP for each computer on the network. In many instance, an ISP will allow you to use a router and connect multiple computers to a single internet connection and pay a nominal fee for each additional computer sharing the connection. This is what when home user will want to look at smaller routers, often called broadband routers that enable you to share two or more computers to share an internet connection."[4]

The basic responsibility of router is to communicate between two networks. So to better understanding of router and his responsibility we should have a good idea about different networks. When we talk about two different net what does that mean, generally when we talk about different networks it means network under different management, but router different networks mean some thing else which we will discuss in next section.

1.2- Different Networks:

As we know that router is responsible to connect different network. It forward he packet of one network to the other network which is know by the router or we can say router has entry about that network in his database or routing tables. When we talk about different networks generally we talk about the networks under different management. It is right for the peoples who does not belongs to computer sciences but for the router different network means networks under different network ID. To better understanding of network ID or different network we have to learn about Internet Protocol (IP). Internet Protocol (IP) or some time Internetworking Protocol (IP) is packet switching protocol. To understand the different we have to learn about IP addressing. Internet Protocol has two versions IPv4 and IPv6. IPv4 is first version of IP (Internet Protocol). Due to lack of addressing space in IPv4 now IPv6 is introduced. On that occasion we don't need to discus detail of IP (Internet Protocol) we will just try to understand addressing of that protocol. And for good understanding we will discus IPv4 addressing.

1.2.1- IP Addressing:

Internet protocol both protocol has different addressing length. IPv4 has 32 bit long address we normally call it IP Address.

IPv4 Address: (32 bits)

As it is very difficult for human to remember 32 combinations of 0s and 1s so IPv4 is represents in decimal form. IPv4 32 bit address is divided into 4 sectors of 8 bits.

Each sector is called octet. These 32 bits are combinations of 0s and 1s, but as we denote it in dot decimal form so final shape of IP address in IPv4 will be.

196.168.1.25

In IPv4 IP address is divided in to 5 classes. Class A, B, C, D and class E. Class A, B and C we use for networks addressing class D and E use for some other purposes like search engine and research purposes. First three classes are use to networks addresses each class has fixed number of networks bits and host bits. Its mean each class has fixed number of networks and hosts. Like in class A first 8 bits are fixed for networks and rest of 24 bits are fixed for hosts. And in class B first 16 bits are fixed for networks and next 16 bits are fixed for hosts and in class C first 24 bits are fixed for networks and last 8 bits are fixed for hosts.

Router just checks the network ID any change in network bits is a new network. So for router different networks means different network IDs. Router knows the network ID with his default network mask. In class full IPv4 addressing number of networks was very less in class A number of networks are 2⁸ and number of hosts in each network was 2²⁴. In class B number of networks are 2¹⁶ and number of hosts in each network are 2¹⁶ and in class C number of networks are 2²⁴ and number of hosts are 2⁸. If we have a close look in class A and B then we will feel in both classes number of hosts having very large space. There is no such network that is having 2²⁴ hosts even 2¹⁶ is also not possible it could be a very rear case. So the idea of classless addressing was introduced to increase the addressing space. The concept of subnetting and supernetting is the part of classless addressing in IPv4. When router receive any packet it just calculate its network ID with is subnet mask and then check its database or routing table to forward it in correct path.

1.3- Introduction to Routing Protocols:

As we know protocol is a set of instructions or rules that controls data, communications and connections. Protocols could be implemented in both ways Hardware and Software and some time combination of the two.[5] Routing Protocols are protocols which are responsible to communicate between routers, to discover the networks either directly connected or remotely connected and to keep updates router about network status. Routing protocols normally work on Layer 3 of OSI (Open System Interconnection) Reference model. In OSI layer 3 is called Networks Layer. OSI (Open System Interconnection) reference model is communication layered model defined by IOS (International Slandered for Organizations). It consists on 7 layers from top to bottom

• Application Layer

• Presentation Layer

• Session Layer

• Transport Layer

• Networks Layer

• Data Link Layer

• Physical Layer

Routing Protocols works on Network Layer. Network layer is responsible for end to end communication. End to end delivery of data packed from source to destination is responsibility of Network Layer. "The Network Layer Performs network routing functions, and might also perform fragmentation and reassembly, and report delivery errors."[6] Network layer use hierarchical logical addressing scheme. Router works on layer 3 and routing protocols also works on that layer.

There are three types of routing protocols Link-State Routing Protocols, Path vector Protocols, and Distance Vector Routing Protocols. I will discus these classifications later on in detail. Routing protocols make routing decision considering some parameters

• Bandwidth

• Delay

• Load

• Reliability

• Hop Count

• Ticks

• Cost

Bandwidth is capacity of a medium to carry data; data capacity of a link is bandwidth. It is very important and basic parameter. Second important parameter is delay; the amount of time that is required to send a packet from source to destination. Load is another parameter, a measure of amount of activity on a route. Reliability of the link should be very height. Cost factor itself consider some parameters and that cost is not only financial cost of the link but also bandwidth and some other factors are also involve.

1.3.1- Routed verses Routing Protocols:

Routing protocols helps the router to take correct routing decision. They discover all network nodes and keep information about all other routers on that network or all connected networks. Routed protocols are responsible to transfer the data between two routers. When router takes any routing decision based on routing protocols then it forward packet to that destination then routed protocol takes that data to that destination. Some routing protocols run over routed protocols like BGP and TCP. So it is important to take care during designing of protocols either routing protocols or routed protocols they should not be depend on each other. Routing protocols runs over particular transport mechanism but some routing protocols have their own transport mechanism. IS-IS (Intermediate System to Intermediate System) runs over data link layer.

RIP run over UDP

BGP (Border Gateway Protocols) run over TCP

OSPF (Open Shortest Path First) has his own transmission mechanism

EIGRP (Enhanced Interior Gateway Protocols) has his own transmission mechanism

IGRP (Interior Gateway Protocols) Unreliable delivery

OSPF, EIGRP and IGRP directly run over IP. "In order to maintain the good service, three key factors have been considered in designing IP Networks: large-bandwidth links, high router data throughput, and high packet forwarding rates."[7]

1.4- Types of Routing Protocols:

Basically there are three types of routing protocols.

Distance vector routing protocols

Link state routing protocols

Hybrid routing protocols

1.4.1- Distance Vector Routing Protocols:

Running a dynamic routing protocol will allow our routers to share their network information with each other and recover from a network outage automatically. If an alternate path exists and all of the routers are running properly configured routing protocols, they will eventually locate the alternate path and use it if the primary path goes down.

Distance vector protocol send information to his neighbor routers. In distance vector protocol all routers send their information to there neighbor router periodically to update their routing table. In distance vector protocol all routers receive information about all networks from their neighbor routers. Distance vector routing protocols has three basic characteristics. Distance vector routing protocols uses the Bellman-Ford algorithm to calculate path.

Updates are frequent and Periodic.

Updates are sent to broadcast address.

Updates are listed almost all known networks.[8]

Distance vector routing protocols are not good for scalable networks. Another problem with distance vector is it requires more bandwidth to send the routing information. Routing loops are also possible in distance vector routing protocols. To avoid that problem there are number of solutions like to set the time to live or hope limit and we can set hold down timer.

We have a simple type of networks let suppose any distance vector routing protocol is implemented in that network. Let see how routers update their routing tables step by step and get information about all connected networks.

Step: 1

Step: 2

In first step all routers learn about their direct connected networks. Routers just make entries of direct connected networks in their routing tables with the information of connected interface and distance. And in step two routers explore networks which are in distance 1. Router keep entry of direct connected networks as distance 0 and the networks which are just one node away are in distance 1 in that way distance vector routing protocols find all networks and update routing tables.

1.4.2- Link State Routing Protocol:

Link-state routing is used with larger networks it's very good for scalable networks. Link-state protocols only send updates of a topology change when any change accrue in network it send that information to all networks so that routers could update their routing table for correct and efficient routing decision. In Link-state Periodic updates are more infrequent than for distance-vector protocols as distance vector send update information periodically it is more likely to coz of usage of bandwidth. Link state just sends the info when any change accrues in networks. Networks running link-state routing protocols can be segmented into area hierarchies, limiting the scope of route changes. Networks running link-state routing protocols supports classless addressing. Each router in Link State Routing must do the following:

• Discover its neighbors, learn their network address.

• Measure the delay or cost to each of its neighbors.

• Construct a packet telling all it has just learned.

• Send this packet to all other routers.

• Compute the shortest path to every other router.

The famous link-state routing protocol is Open Shortest Path First (OSPF). OSPF developed by Internet Engineering Task Force (IETF). OSPF is classless, has a very fast convergence time and is complicated, both in operation and in configuration, just as one would expect from something developed by a committee. OSPF is open protocols so specification is public domain. Link-state protocol convergence time is very fast then distance vector protocol; distance vector is slow in convergence time because of loop avoidance feature. Loops are not possible in link-state protocols that is built feature. Link-state required more memory and CPU usage then distance vector.

1.4.3- Path Vector Routing Protocols:

Path vector routing protocols are different from distance vector and link state routing protocols. BGP is based on a routing method called path vector routing. It is different from both distance vector and link-state routing. AS boundary Routers (ASBR) that participates in path vector routing advertise the reach ability of networks in their own AS to neighbor ASBRs. Two ASBR are considered to beneighborsof each other if they are connected to the same network An ASBR receives its information from an interior routing algorithm such as RIP or OSPF

Each Entry in the routing table contains:

The destination network

The next router

The path to reach the destination

1.5- Routing Protocols:

In this section we will discuss and compare some dynamic routing protocols belongs to each category. We will compare following routing protocols:

Routing Information Protocol (RIP)

Interior Gateway Routing Protocol (IGRP)

Enhanced Interior Gateway Routing Protocol (EIGRP)

Open Shortest Path First (OSPF)

Intermediate System to Intermediate System (IS-IS)

1.6- TCP/IP Model:

TCP/IP stands for Transmission Control Protocol - Internet Protocol is a protocol based on two basic protocols TCP and IP. TCP/IP is layered model consist on four layers.

• Network Layer

• Internet Layer

• Transport Layer

• Application Layer

TCP/IP layers based on OSI (Open System Interconnection) layers. Network layer is responsible for packet transmission and reception. It depends on topology used. TCP/IP could be used both reliable and unreliable transmission. It used two protocols TCP (Transmission Control Protocol) and UDP (User Datagram Protocol).

OSI vs TCP/IP

1.7- Label Switching:

Label switching is a technique to achieve the reliable delivery of packet in data link layer rather then network layer. Label switching is an advance approach to get the reliable delivery for advanced broadband networks. MPLS (Multiprotocol Label Switching) is also use label switching technique. "MPLS is a connection oriented datagram forwarding technique based on short labels which are assigned to paths and to packets forwarding on these paths."[9] Label switching technique allows the packet switching on the bases of labels. Each node in the network contains a database of labels assigned to the packets. It takes to switching decision on the bases of that label database. It reduces the time of routing decision. MPLS uses the label switching technique but the reason MPLS is getting famous is that it also support QoS with label switching. MPLS use the same label switching technique which ATM switches use. Label switching technique will speed up internet routers dramatically. It improves the performance of back bone routers. One of the basic reasons to achieve that speed is the limitation of packed size and MTU (Maximum Transmission Unit) will give that limitation.

2- Introduction to dynamic routing Protocols:

The concept of dynamic routing is very important for today's networks and dynamic routing play very important role. Because today's networks requirement changing constantly and in that situation static routing is not good idea. In the case of static routing after each change in any part of network we have to change configuration of all dependent part of that network. Dynamic routing solves that problem to exchange of topology state information and keeping updates the network nodes about networks status, so that networks nodes can make correct routing decision. In the case of WAN (Wide Area Networks) each node of network may have more then one path any destination but there must be one path which will be the best path among all others. Best path mean the path in less cost. It could be the shortest path as well but that is not important less cost is important. Dynamic routing algorithms found that path and forward traffic on that path.

2.1- Evolution of Dynamic Routing Protocols:

"Dynamic routing protocols have been used in networks since the early 1980s. The first version of RIP was released in 1982, but some of the basic algorithms within the protocols were used on the ARPANET as early as 1969."[10] The ARPANET (Advance Research Projects Agency Networks) was the first packet switching network developed by ARPA. Internet is a latest form of ARPANET we can say ARPANET is a predecessor of internet. The routing protocol RIP (Routing Information Protocol) was the first routing protocol. RIP (Routing Information Protocol) used Bellman-Ford Algorithm which was developed in 1968.[11] Later on a new version of RIP protocol was introduced RIP2. But both RIP and RIP2 was not suitable for large scale networks. For large scale networks we need some advances routing protocols then OSPF (Open Shortest Path First) and IS-IS (Intermediate System-to-Intermediate System) was introduced. Later on CISCO developed two protocols for large scale networks Interior Gateway Routing Protocol (IGRP) and Enhanced IGRP (EIGRP).

Table No. 1[12]

2.2- Role of Dynamic Routing Protocols:

The basic role of dynamic routing protocol is to update network nodes about the status of network atomically. Routing protocols updates routers and make them able to get the information about other routers or network nodes which are connected remotely. Every router belongs to any network save entry about every other router connected to the same network in his routing table. Either these routers are connected directly or remotely.

We use dynamic routing protocols to reduce the administrative overhead. If we use static routing when ever we need to change any node or we need to add any node in network we have to reconfigure the network. In other case if any node of network is not working properly and we have to change it or remove it even then we have to reconfigure many part of network. It is very important for router to keep update routing table about the status of network for the correct routing decision.

Although dynamic routing is more appropriate solution but we cant ignore the advantage of static routing. If in some time dynamic routing is performing good, but some other time static routing still has its place. Mostly time we need both types together some part of network required dynamic routing and some part required static routing. Both types have their own advantages and disadvantages.

2.3- Network Discovery:

The purpose of dynamic routing protocol is not only to make a correct routing decision for the better communication it is also responsible for some other issues and Network discovery is one of them. Network Discovery is responsibility of dynamic routing protocol it discover all direct connected networks and also remotely connected networks. Dynamic routing protocols not only discover the networks it also keeps update about network status.

Basic purpose of routing protocols is to choosing the best path to destination networks. It should have the ability to find the new best path if the current path is not longer available. Some time due to any reason one path for a particular destination could block or may be it no longer exists. Routing protocols should have ability to find another path for that destination. Protocols do that with the ability of discovering networks. Routing protocols are able to maintain up to date routing information and they find some part of networks is not in working condition may any one node is no longer exists, so routing protocols found any alternative path for destinations depending on that node. Routing protocols use messages to discover remote networks, maintain up to date routing information and choose the best path to destination. IT exchange information in messages and keeps that information in tables or uses some databases.

2.4- Dynamic routing in Telephone Networks:

In switching networks there are two major types of networks first one is Circuit Switching Networks and second is Packet switching networks. In traditional telephone networks we use circuit switching networks, in circuit switching when two nodes want to communicate they establish a connection between them before they start communication. Both nodes establish a connection with signalling and all communication done to fallow this path. Circuit switching is fastest switching technique because they use full bandwidth of selected path, path is not sharable. And second technique is packet switching, in packet switching data travel in packets and in each node packet decides his route. So routing in packet switching is difficult then routing in telephone networks. A PSTN (Public Switched Telephone Networks) use a fully meshed topology or a strongly meshed topology. PSTN always use hierarchical networks structure. There are many routing scheme which PSTN can use some of them are:

• Dynamic Non-Hierarchical Routing (DNHR)

• Real Time Network Routing (RTNR)

• Dynamic Alternative Routing (DAR)

• Dynamically Controlled Routing (DCR)

• Dynamic Routing 5 (DR5)

• Intelligent Network Based Dynamic Routing (IN/DR)

• Trunk Status Map Routing (TSMR)

• System for Testing Adaptive Traffic Routing (STAR)

• Forward Looking Routing (FLR)

• State Dependent Routing (SDR)

• State and Time Dependent Routing (STR)

2.5- Classification of Dynamic Schemes

2.5.1- Hop-by-Hop Routing:

In hop-by-hop routing method routing decision is taken by each node in the path. This method is implemented in circuit switched telephone networks. And now this technique is also widely used in data networks. In data networks each node in the path review routing table and take routing decision for that packet. In data networks there are some chances to occur routing loops but in telephone networks routing loops are avoiding by hierarchical routing. In data networks routing loops could occur so we use a field in header which set the maximum number of nodes that packet can travel and if in that particular number packet could not reach to destination packet discord by any node. The name of that field is Time-to-live when any node receive a packet before it compute his path it check its time to live field if it reaches to maximum limit it discord that packet and consider that it was lost packet.

2.5.2- Source Routing:

There are two way of communication in data networks connection oriented and connectionless. Hop-by-hope routing method is use in connectionless routing. In the case of connection oriented, source to destination connection are established before forwarding the packets and all packets use same path to destination. When communication is completed we terminate the path that method is called source routing. In source routing source node establish the connection and then start communication and when communication is completed terminate the connection. Source routing in ATM networks use both MPLS (Multiprotocol Label Switching) and PNNI (Private Network Node Interface).

2.5.3- Hierarchical Routing:

To arrange the routers in hierarchical manners is called hierarchical routing. We use hierarchical routing to avoid the routing loops. TCP/IP based hierarchical routing in TCP/IP there are two level one is network level and second is host level. From gateway to gateway routers only use network portion of address for routing decision. After gateway it use host portion of address to find the particular end user. "The concept of hierarchical routing is used in telephone networks in order to avoid routing loops in large networks, as well as in PNNI networks, for scalability and networks configuration purpose."[13]

2.6- Degree of Meshing:

Routing algorithms compute the best path between two nodes. To compute the best path first we should know how many paths available between these two nodes so we can say that degree of meshing is very important for routing algorithms. Degree of meshing determines that how many paths are available and then algorithm applies to find the best one. Degree of meshing also determines the complexity of algorithms. If the number of paths will increase complexity of algorithm will increase.

2.6.1- Fully Meshed:

In fully mashed topology every node is directly connected to every other node. We can say that is a complete topology because in fully meshed every node got point to point connection to every other node. In fully meshed network each node can send data each other node simultaneously. But the main disadvantages of those networks are that it's very costly.

2.6.2- Partially Meshed:

In partially meshed networks each node is connected more then one of network but not all other nodes. Each node have limited point to point link to other nodes. In partially meshed topology strong meshed and weak meshed are very user full in strong meshed and weak meshed topology some of link from fully mesh networks are remove. If we delete some of link from fully mesh network we will obtain strongly mesh or weakly meshed networks. The difference between strong mesh and weak mesh is according to scalability factor.

"In graph theory, the number of links originating at a network node is called the degree of the node. With the above nation of strong and weak mesh topologies it result that the average degree of nodes increase asymptotically as the number of nodes for strongly meshed networks, whereas it remains constant for weakly meshed networks"[14]

.

The mesh degree of a networks could be different at physical layer and it could be some thing else in other layers. Means the degree of mesh in networks at networks layer or higher layer then physical depends on protocols so it may be different then physical layer. In ATM networks, Physical links are logical divides in Virtual Circuits (VC) and Virtual Paths (VP) and Virtual Path Connections (VPC). Because in ATM we configure VPC's and one VPC contains more then one VP's and every VP contains more then one VC's. So if we will configure new VPC in ATM network then degree of mesh will be change.

3- Routing Computation Algorithms

In this section we will discuss routing computation algorithms. Some routing algorithms are quit simple but for better understanding of routing algorithms we should have idea about graph theory. Because in examples router will be known as nodes and links will be known as edges.

3.1- Dijkstra's Algorithm:

Dijkstra's algorithm works as forward search theory. E. Dijkstra a Dutch scientist was the founder of this algorithm. Dijkstra's algorithm founds lost cost paths from a single source to all other nodes. It works only from single source and find least cost paths to all other nodes in networks.

Working of Dijkstra's algorithms is quit simple. We will start from source node and make a list of distance for all other nodes from source node and set distance as infinity except source list, distance of source list will be zero. And set all the value in visited list as false. Set all the values in previous as undefined.

1. Create 4 lists distance list, current node list, previous node list, visited node list.

2. Set all the values in distance list as infinity except source node because distance from source to source will be zero.

3. Set current node as starting nodes.

4. Change the value of current node as visited node.

5. Update the distance value based on previous node to current node which is directly connected.

6. Set the current node as reachable from source in shortest path and also mark it unvisited.

7. Repeat from step 4 until all nodes are explored.[15]

In figure 10 if we apply Dijkstra's algorithm considering B as an source node. Then the iteration will be.

3.2- The Ford-Fulkerson Algorithm:

Ford-Fulkerson algorithm is also known as Backward Searching algorithm. Ford-Fulkerson algorithm was developed in 1956 for the maximum flow networks. Ford-Fulkerson algorithm finds the least cost rout from any node to destination node. We can its opposite of Dijkstra's algorithm in Dijkstra's algorithm finds least cost rout from source to every other node and Ford-Fulkerson algorithm finds least cost rout from every node to destination node Ford-Fulkerson algorithm is also iterative algorithm but start from destination node.

The algorithm consists of only few steps We need maximum N-1 steps to explore N-1 nodes for N number of node. D is representing destination and C is cost of link. Initially we need to set C (D) = 0 C (N) = infinity.

"For all nodes n sets C (n) = Min {C (W) + l (n, W): W is node in the network}, Where l (n, n) = 0.

The iteration is repeated until no more changes occur"[16]

3.3- The Bellman-Ford Algorithm:

Bellman-Ford algorithm is also a single source algorithm. Bellman-Ford algorithm computes the shortest path of all other nodes in networks from a single node. It computes minimum cost path. It is also iterative algorithm. In any iteration it computes the all edges from the source node to a particular level. In next iteration the level increases. For example Bellman-Ford algorithm starts from source node and in first level it check all directed connected links and check which nodes it can explore in that level and mark them. In next step it check the links of second level not only all direct connected nodes of source but also direct connected nodes of nodes which it explore in first iteration. And during that step if it found any node which is already know by first step if second path is minimum cost path then first one then it will update that path other wise it will remain as it is. This process will repeat until the result will be same in last two iterations.

If N is set of nodes in network, s is source node, h is number of hope being consider, d_ij is distance from i to j and D_n is cost of path from s to n with no more than h hopes then algo will be. Set i = 1;

1. Initialize D_n⁰ = infinity, for all n ≠ s.

2. Find all nodes 1 hope away.

3. Find all nodes i hops away.

4. i = i +1

5. Keep the path for which the cost is minimum

6. Iterate step 3,4 and 5 until the result for all paths remain same.

3.4- Multidimensional Link Metrics:

Multidimensional cost is a link cost which based on more then one aspect of cost. Multidimensional consider many issue to compute the cost of any link. When we compute the cost in Multidimensional link metric we consider number of hops, bandwidth, cell loss ratio, delay to transfer the cell and traffic load. In Multidimensional link metric we also observe the situation of all aspects in different time scale. Multidimensional algorithm based on Dijkstra's algorithm and it needed for ATM traffic.

3.5- Minimum Hope Routing:

In minimum hope routing we only consider number of hops as metric. It is very simple routing technique but it is very famous in data networks. The way to find the path in minimum hop routing is very simple the less number of hopes between sources to destination is shortest path. Many routing protocols are using minimum hope routing as path computing metric. The less number of hopes is best path in minimum hope routing. It does not consider the costs of link just number of hopes are considered as a metric.

If we want to select the path from node A to node G with minimum hope routing then we got two paths first A-D-G and second A-B-G. We can any of these two paths but not any other. We have some other paths from A to G.

• First path is A-B-G which is one hope away.

• Second path is A-B-C-G which is two hops away

• Third path is A-D-G which is one hope away.

• Fourth path one is A-D-B-C-G which is three hops away.

• Fifth one is A-D-B-G which is two hops away.

• Sixth path is A-E-F-G which is also two hops away.

• Seventh path is A-E-F-D-G which is three hops away.

So path first and third is minimum hope routing path so we will select any of them.

3.6- Minimum Cost Routing:

Minimum cost routing is depend on the cost of the link. For minimum cost routing we should know the cost of each link. The cost of link is computed by how frequently that link being used if link is using very frequently it cost will be high. For example we got two links link A and link B. Link A is being used once in a min and link B used 4 or 5 time in a min. The cost for the link may be 1 and the for the link B will greater then link A it may be 2 or 3. So in minimum cost routing it always prefers the link with low cost in that example we will consider link A is best then link B because its cost is 1 and that is good then cost of B which is 2 or 3.

If we want to chose the path between A and G we will chose the path which link cost will be minimum. We not consider the number of hopes in that case first of all we should know how many path available between A and G.

• First path is A-B-G with the link cost4.

• Second path is A-B-C-G with the link cost6.

• Third path is A-D-G with the link cost6.

• Fourth path one is A-D-B-C-G with the link cost8.

• Fifth one is A-D-B-G with the link cost6.

• Sixth path is A-E-F-G with the link cost6.

• Seventh path is A-E-F-D-G with the link cost3.

So the best path according to minimum cost routing is path A-E-F-D-G because its total link cost is minimum. According to minimum hope routing it's a longest path but according to minimum cost routing its a shortest path between A and G.

4- Comparison of Different Routing Protocols

4.1- RIP (Routing Information Protocol):

Routing Information Protocol (RIP) is the first routing protocol up till now RIP has three versions RIPv1, RIPv2 and RIPng. RIP is a interior gateway routing protocol. It is classified as network layer protocol it's also run on top of UDP. RIP has limited broadcast routers on same segment exchange updates through broadcasts. RIP is a distance vector routing protocol that sends updates every 30 seconds. After 30 second it sends its entire routing table, whether or not changes have occurred within the network. RIP uses a single metric in determining the best path, hop count. The maximum number of hops is 15; anything over that is invalid and indicates an unreachable network. Every routing protocol determines his path on the bases of metric. In RIP metric is Hope Count; less number of hops means best path.

If we implemented RIP protocol in figure 1 network. To understand the hope count we will just discuss one path. Let suppose the networks in figure 14 is implemented with RIP protocol. In router A it got entry of router D with the path of A-B-D because RIP metric is hope count. A got 4 options A-B-D, A-C-D, A-C-E-D and A-C-E-F-D and first two options are same in distance of 1 other two options are in the distance of 2 and 3. RIP could choose any one from first two options.

"End hosts or routers may implement RIP to keep track of routes to destinations, such as hosts, networks, subnets, and default routes. End hosts or routers learn routes to destination via routes updates that device exchange dynamically on the local segment via periodic broadcasts. Although end hosts are capable of running RIP, they generally do not, instead relaying on routers to keep track of rout information. When RIP is enable on an end host or router a local UDP port is activated for the routing process. All devices running listen on UDP. Port 520 and transmit from the same UDP port"[17]

4.1.1- RIP Version 1 Header:

Table 3:RIP Header[18]

Command:

Command is 1 byte field that is identified the purpose of that packet. There are eight types of commands.

Version:

Version field identifies the RIP version. RIP has two versions version 1 and version 2 both versions are compatible. Eight bits of version field identifies which version is in use.

Unused:

There are many fields in RIP header which are unused.

Address Family Identifier:

That field always contain value 2 that represent to IP. May be RIP support some other protocols in future that's why that field is added in RIP header.

IP Address:

That contains 32 bits IP address. Different between RIP version 1 and RIP version 2 is RIP version 2 support classless IP address but RIP version 1 does not support classless IP. So RIP version 2 also called classless RIP. Latest version of RIP is RIPng which is design for IPng protocol also called IPv6.

Metric:

In RIP metric is hope count so that field keep the information of number of hopes. RIP has maximum hope limit 15, so the value of that field must be between 1 to15.

4.1.2- Disadvantages of RIP version 1:

RIP is one of first routing protocols and it is very popular many networks are still using RIP but it has some disadvantages. RIP is a broadcast based protocol it broadcast his updates to all networks and that is coz to use of bandwidth. RIP is distance vector routing protocol so it sends it updates periodically. RIP send its entire table to all networks even when no changes occur. Metric of RIP is hope count and its only support up to 15 hopes. And RIP is a classful routing protocol so does not support VLSM (Variable Length Subnet Mask).

4.1.3- RIP Version 2:

RIP version 2 was designed to over come some of limitations in version 1 and also enhanced some features. RIP version 2 is not classful it is enhanced to support VLSM. "With more intelligent routing protocols and the inherit limitation of RIP, version 2 has been overlooked in most corporate network implementation."[19]

Table 4: RIP Version 1 versus Version 2[20]

We can see even in RIPv2 limitation are still there the maximum hope count limit is still 15 hopes. But version 2 is classless that is big advantage.

4.2- Interior Gateway Routing Protocol (IGRP)

The Interior Gateway Routing Protocol (IGRP) is a dynamic distance-vector routing protocol. IGRP designed by Cisco in mid 1980's for routing with in an autonomous system that contains large, arbitrarily complex networks with diverse bandwidth and delay characteristics. The definition of autonomous system is a collection of network under a common administration. CISCO gave several routing protocols. IGRP is a classful, distance vector routing protocol similar to RIP version 1. IGRP does not include network mask information. As we discuss it is distance vector routing protocol so it sends its periodic updates, and its periodic updates occur every 90 seconds by default. IGRP updates have a destination address of the IP broadcast address configured on the interface out which the updates are being transmitted; the default IP broadcast address is 255.255.255.255. IGRP was implemented only to work in Internet Protocol IP networks.

As we discuss IGRP initially work on broadcasts so for initialization it send route request to all neighbor routers and broadcast their entire routing table. All routers do the same broadcast and exchange their routing tables and update routing entries. And when they maintain all routing entries for entire network they keep updating their information by broadcasting entire routing table after every 90 second. The metric used by IGRP is a composite (calculated value) of five parameters that are carried in the update packets.

The five parameters are as follows:

• Bandwidth

• Delay

• Reliability

• Load

• MTU

Bandwidth:

The higher the bandwidth is represents the better path. It is a configurable value that is represents the transfer rate specified in kbps and is configured with the interface configuration commandbandwidth.

Delay:

The links have less end to end delay is better then the link have more delay. The delay is supposed to be the amount of time a bit takes to cross a network.

Reliability:

A value shows the reliability of a link in percentage. If the reliable value is higher then the path is better. The more reliability shows the better path.

Load:

Load of traffic in a link is means how often that link is being used. It represents the utilization of any link. The lower load value is represents better path.

All of these parameters are tracked by IOS for each interface in a router. Bandwidth and delay are the two metrics that are commonly used and they comprisedefault metric. Bandwidth and delay are static; they have default values, and the values are configurable.

Configured with the interface configuration command

A(config)# interface serial 0

A(config-if)# bandwidht 56

This bandwidth 56command in this example tells IOS that Serial0 is connected to a 56 kbps WAN. If the bandwidth setting on an interface is not the same as the actual bandwidth of the directly-connected network, you should use the bandwidthcommand to change the setting.

RIP does not support multiple autonomous systems. IGRP has the capability of creating different autonomous systems its working in not like OSPF autonomous system. When we run IGRP, we give it an autonomous system number. If this number is kept the same, then all the routers comes within the same autonomous system. The default administrative distance for IGRP is 100. If we were running both RIP and IGRP in our internetwork, IOS would prefer the routes learned from IGRP because IGRP's administrative distance is lower than RIP's. RIP's administrative distance is 120.

4.2.1- Interior Gateway Routing Protocol (IGRP) Features:

The important IGRP characteristics are the more scalability than RIP. IGRP scalability feature is very high then RIP. And second important feature is its periodic updates. It sent periodic updates after 90 second where RIP sends his periodic updates after every 30 second so its three time less overhead in that case. The third important feature is fast response to network changes. IGRP path selection is very sophisticated. It considers 5 different parameters; Bandwidth, Delay, Reliability, Load and MTU as metric to calculate path. And it supports Multiple-paths.

4.3- Enhanced Interior Gateway Routing Protocol (EIGRP):

EIGRP is enhanced version of interior gateway routing protocol. Same as IGRP, EIGRP is also distance vector routing protocol with enhanced link-state characteristic. Metric calculation of EIGRP is same as IGRP. It also considers all five parameters to calculate metricBandwidth:Bandwidth is a configurable value that is represents the transfer rate specified in kbps and is configured with the interface configuration commandbandwidth. The higher the bandwidth is represents the better path. Second isDelay:Delay is measure the time it takes in queue in a router specified in microseconds and is configured with the interface configuration commanddelay. The links have less end to end delay is better then the link have more delay. The delay is supposed to be the amount of time a bit takes to cross a network. The third one isReliability:A value shows the reliability of a link in percentage. If the reliable value is higher then the path is better. The more reliability shows the better path. EIGRP supports Rapid convergence and it reduce the usage of bandwidth and the most important point of EIGRP is it supports multiple network-layer protocols.

4.3.1- EIGRP Features and Advantages:

One special feature of EIGRP is it supported multiple upper layer protocols, such as IPX, IP and AppleTalk. Network running in more then one protocol suit normally required one protocol for each protocol suit, but EIGRP a single protocol can be used to build routing information. EIGRP is an advanced distance vector routing protocol with enhanced link-state characteristic. Another important characteristic of EIGRP is that it is 100% loop free routing protocol and it gives fast convergence and easy configuration. EIGRP is very easy to configure its configuration is quit simple and very easy to understand. Another important feature of EIGRP is it got less network design constraints than OSPF. EIGRP sends Incremental updates. And it supports Variable Length Subnet Mask (VLSM) and discontinuous networks it also support classless routing. RIP version one does not classless routing. EIGRP is fully compatible with existing IGRP networks and it supports more then one upper layer protocols like IPX and AppleTalk.[21]

The main advantage of EIGRP is it uses multicast instead of broadcast and that reduce the overhead on links. And second important aspect is it considers link bandwidth and delay in metric to calculate paths. EIGRP also perform unequal cost path load balancing and it is more flexible than OSPF. As we mention EIGRP is an advanced distance vector routing protocol so it automatically establishes neighbor relationships with peer devices. EIGRP relies on IP packets for delivery of routing information.

EIGRP supports

• Multi access (LANs)

• Point-to-point (HDLC)

• NBMA (Frame Relay)

• Variable-length subnet masks (VLSMs)

• Hierarchical designs

EIGRP use five types of packet in operations

• Hello: It establish the neighbor relationships

• Update: Send routing updates

• Query: Ask neighbors about routing information

• Reply: Response to query about routing information

• ACK: Acknowledgement of a reliable packet

4.4- Open Shortest Path First (OSPF):

Open Shortest Path First (OSPF) is a routing protocol developed for Internet Protocol (IP) networks by Internet Engineering Task Force (IETF) in 1989. OSPF was created because in the mid-1980s, the Routing Information Protocol (RIP) was gradually more incompetent of serving large, various internetworks. OSPF version 1 was an experimental protocol built for routers and UNIX work stations. It was experimental protocol so it never been implemented. OSPF version 2 was introduced in 1991 with many technical improvements. It is the most commonly used protocol in the world. OSPF is an open standard, link state routing protocol. OSPF was developed by the Internet Engineering Task Force (IETF), and its latest version, 3, is defined and version 3 is for IPv6. OSPF is classless, has a very fast convergence time and is complicated, both in operation and in configuration, just as one would expect from something developed by a committee.

OSPF is a link state routing protocol and as we discus earlier that link state routing protocol has many advantages in distance vector routing protocol. Link-state routing is used with larger networks as Link-state protocols only send updates of a topology change. In link state protocols periodic updates are more infrequent than for distance-vector protocols. Networks running link-state routing protocols can be segmented into area hierarchies, limiting the scope of route changes. Link-state routing protocols supports classless addressing. Each router in Link State Routing must discover its neighbors, learn their network address. It measures the cost and delay to each of its neighbors. It constructs a packet telling all it has just learned and send this packet to all other routers. Compute the shortest path to every other router.

4.4.1- Characteristics of OSPF:

Open Shortest Path First OSPF has two primary characteristics. The first is that the protocol is open, which means that its specification is in the public domain. The OSPF specification is published as Request For Comments (RFC) 1247. And the second principal characteristic is that OSPF is based on theSPF algorithm, which sometimes is referred to as the Dijkstra's algorithm, named for the person credited with its creation. OSPF can operate within a hierarchy. OSPF a hieratical based protocol. It works on different areas and different autonomous system AS. The largest unit within the hierarchy is the autonomous system (AS), which is a collection of networks under a common administration that share a common routing strategy. OSPF is an intra-AS (interior gateway) routing protocol, although it is capable of receiving routes from and sending routes to other AS's. OSPF supports classless routing.

OSPF share its routing advertisement in multicast rather then broadcast that reduce a lot of overhead. In OSPF, the AS Number is identified by ID that is called process ID. Only routers with same OSPF process ID can communicate to each other. In OSPF AS could be divided into numbered areas, where an area is a network or a set of closest networks. Areas do not overlap, but it may be possible, that some routers in an AS belong to no area. And it is also possible some router belongs to two areas these routes called Area Border Router (ABR). An area is a simplification of a subnet. An area is created to control the information (LSAs) by your own will. A router belong to an area only have information of that area outside an area, its topology and details are not visible. Within an area, each router has the same link state database and runs the same shortest path first algorithm. Its job is to calculate the shortest path from itself to every other router in the area.

4.4.2- Types of Router in OSPF:

There are four types of routers are used in OSFP environment

• Internal Routers

• Backbone Routers

• Ares Border Routers (ABR)

• Autonomous System Boundary Routers (ASBR)

Internal Routers:

A router who's all interface belongs to same area is called internal router. Internal routers are within an area

Backbone Routers:

A router which belongs to the backbone area is called backbone router. OSPF routing limited to backbone.

Ares Border Routers (ABR):

Area Border Routes connect two or more areas. A router which connects one area to other area is called ABR. These routers maintain separate topological databases for each area. They run the shortest path first algorithm for each area separately.

AS Boundary Routers (ASBR):

A router which connects two autonomous systems called ASBR. An AS can be divided into a number of areas, which are groups of contiguous networks and attached hosts.

An OSPF backbone is responsible for distributing routing information between different areas. It consists of all Area Border Routers, networks not wholly contained in any area, and their attached routers. The backbone topology is invisible to all intra-area routers, as are individual area topologies to the backbone. Every AS has a backbone area, called area 0. All areas are connected to the backbone, possible by tunnels (virtual paths). Thus, it is possible to go from any area in an AS to any other area in the same AS via backbone. A tunnel is represented in the graph as an arc, and has a cost. Each router that is connected to two or more areas is a part of the backbone. The topology of the backbone is not visible outside the backbone area.

4.4.3- OSPF Metric:

OSPF uses a single metric calledcost.The cost is calculated from the bandwidth assigned to an interface; this bandwidth is defined the same way as the bandwidth used in metric for IGRP and EIGRP. The cost of a link is included in the LSA that is being used to advertise the link.

4.4.4- Advantages of OSPF on RIP:

OSPF is going to converge quicker then RIP protocol. The reason is RIP is a distance vector protocol and distance vector protocol sent periodic updates; in case of RIP protocol RIP send updates of each 30 second. RIP send his update after 30 second and to his neighbours and it sent entire table to all neighbour nodes. OSPF only sent updates when there is some change in networks status and it only sent changes to all nodes not entire table.

A second main advantage of OSPF is that OSPF supports VLSM (Variable Length Subnet Mask). Another advantage of OSPF on RIP is size. OSPF is unlimited there is no hard and fast size for OSPF. The size of RIP protocol is 15 hopes. RIP can work till maximum 15 hopes. The fourth advantage of OSPF is it uses less bandwidth in signalling. As we discuses that RIP sent entire table after each 30 second so it use more bandwidth then OSPF which only sent changes to neighbour router.

Metric is another advantage of OSPF. OSPF consider bandwidth and some other factor as metric and metric is called cost. RIP only consider hope count as metric.

4.5- Intermediate System - to - Intermediate System (IS-IS):

The Intermediate system to intermediate system (IS-IS) is a link-state routing protocol developed by ISO (International Standards Organization). IS-IS is an interior gateway routing protocol as we discus IGP in routing types. Interior Gateway Protocols works with in autonomous system. And another type Border Gateway Protocols works between autonomous systems (AS). IS-IS was developed to work with Classless Routing Protocol (CLRP).

IS-IS is link-state routing protocol so it's mean it use flooding to exchange the routing information. IS-IS is a popular protocol in Digital Equipment Corporation DECnet. Each version of DECnet protocols is known as Phase I, Phase II, Phase III, and so on. IS-IS is DECnet phase V. "IS-IS was developed at roughly the same time that the Internet Engineering Task Force IETF was developing a similar protocol called OSPF. IS-IS was later extended to support routing ofdatagrams (aka network-layer packets) using IP Protocol, the basicrouted protocol of the global (public) Internet. This version of the IS-IS routing protocol was then called Integrated IS-IS (RFC 1195)."[22]

IS-IS is very similar to OSPF (Open Shortest Path First) both are link state routing protocols and both use Dijkstra's algorithm to calculate shortest path. IS-IS also support VLSM (Variable Length Subnet Mask) like OSPF and IS-IS also use multicast to discover neighbour routes like OSPF. OSPF use multicast to discover his neighbour router and it hello packet to do that same as OSPF, IS-IS also use hello packet to discover his neighbour routers. But the difference between OSPF and IS-IS is OSPF is IP based routing protocol and work on layer 3. IS-IS does not work with IP it does not use IP to carry data. And IS-IS works on network layer as CLRP works on network Layer.

IS-IS is a hierarchical routing protocol. It works on different areas and domains. IS-IS recognise his devices by areas and domains. It divides networks in areas and domains so for administrator it is very easy to control the flow of data.

4.6- Border Gateway Routing Protocol (BGP):

TheBorder Gateway Protocol (BGP) is an inter-autonomous system routing protocol. BGP use to communicate between two autonomous systems (AS). It is a path vector routing protocol. An autonomous system is a network or group of networks under a common administration and with common routing policies. There are some disadvantages in distance vector routing protocol that's why path vector routing protocol use for inter autonomous system. Distance Vector is not a suitable protocol for smallest hope count because many times the route with the smallest hop count is not the preferred route. May be this smallest hope count could not give us enough bandwidth and we got another route which is not smallest hope count but more then enough bandwidth with less delay so we should go for that. For example, we may not pass through an AS that is not secure even though it is the shortest. Distance Vector routing is unstable due to the fact that the routers announce only the number of hop counts to destination without actually defining the path that leads to that destination. For example a router that receives a distance vector advertisement packet may be fooled if the shortest path is actually calculated through the receiving router itself.

There is also a drawback in link-state routing protocols. Link state routing in also not a good for inter-autonomous system routing because an internet is too big for this routing method. Link state routing for the whole internet would require each router to have a huge link state database. It would also take a long time for each router to calculate its routing table using the Dijkstra algorithm. BGP is used to exchange routing information for the Internet and is the protocol used between Internet service providers (ISP). Customer networks, such as universities and corporations, usually utilize an Interior Gateway Protocol (IGP) such as RIP or OSPF for the exchange of routing information within their networks. Customers connect to ISPs, and ISPs use BGP to exchange customer and ISP routes.

There are two types if BGP routing protocol

• External Border Gateway Protocol (EBGP)

• Internal Border Gateway Protocol (IBGP)

External Border Gateway Protocol (EBGP)

When BGP is used between autonomous systems (AS), the protocol is referred to as External BGP (EBGP).

Internal Border Gateway Protocol (IBGP)

If a service provider is using BGP to exchange routes within an AS, then the protocol is referred to as Interior BGP (IBGP).

For the first time BGP neighbors exchange full routing information when the TCP connection between neighbors isestablished. But the full routing information they exchange only for first time when connection established. When connection established if changes to the routing table are detected, the BGP routers send to their neighbors only those routes that have changed. BGP routers do not send periodic routing updates, and BGP routing updates advertise only the optimal path to a destination network. BGP is based on a routing method called path vector routing It is different from both distance vector and link-state routing. In path vector each entry in the routing table contains the destination network, the next router address and the path to reach the destination. Autonomous System Boundary Routers (ASBR) that participates in path vector routing advertises the reach ability of networks in their own AS to neighbor ASBRs. Two ASBR are well thought-out to be neighborsof each other if they are connected to the same network. An ASBR receives its information from an interior routing algorithm such as RIP or OSPF. Each router that receives a path vector message verifies that the advertised path is in agreement with its policy and if it is, the router updates its routing table and modifies the message before sending it to the next neighbor. The modification consists of adding it's AS number to the path and replacing the next router entry with its own identification.

BGP Messages:

BGP use four types of messages and it exchanged messages using TCP.

• OPEN:Opens TCP connection to peer and authenticates sender

• UPDATE: Advertises new path (or withdraws old)

• KEEPALIVE:Keeps connection alive in absence of UPDATES it also ACKs OPEN request

• NOTIFICATION:Reports errors in previous messages and also used to close connection

To create a neighborhood relationship, a router running BGP opens a TCP connection with a neighbor and sends an open message. If the neighbor accepts the neighborhood relationship, it responds with a keepalive message, which means that a relationship has been established between the two routers. The routers running the BGP protocols, exchange keepalive messages regularly before their hold time expires. This message tells other routers that they are alive or there status is on working. Keep alive message consists of only the common header of 19 bytes of BGP packet.[23]

5- Asynchronous Transfer Mode

5.1- ATM Technology:

ATM stands for Asynchronous Transfer Mode. Asynchronous Transfer Mode is a streamlined packet transfer interface. Asynchronous Transfer Mode (ATM) is a standard given by International Telecommunication Union-Telecommunications (ITU-T). ATM is similarities to packet switching networks. An ATM transfer data in separate chunks and is supports multiple logical connections over a single physical interface. It means ATM divides a physical medium into many logical connections. A single physical path could have many logical paths in it.

A ATM physical connection is divided into three logical connections. ATM uses fixed sized packets called cells that's why ATM also known as cell relay. ATM is a smooth packet transfer interface which is very reliable and devotion of modern digital facilities to provide faster packet switching with minimal error and flow control and it achieves data rates of 25.6Mbps to 622.08Mbps.

In ATM, information for multiple service types, such as voice, video, or data, is communicate in small, fixed-size packets called cells. ATM agreed as transfer mode for Broadband ISDN. ATM is asynchronous transmission it means no exclusive reservation of resources to calls and allocation of resources are on need and priority basis. It uses statistical multiplexing. Similar to other packet switching and frame relay, ATM involves the transfer of data in distinct chunks, and allows multiple logical connections to be multiplexed over a single physical interface. In the case of ATM, the information flow on each logical connection is organized into fixed-size packets, called cells.

5.2- ATM Reference Model:

ATM reference model involves three separate planes:

• User Plan

• Control Plan

• Management Plan

User Plane:

ATM user plan provides for user information transfer, along with associated controls, flow control and error control.

Control Plane:

Control plan performs call control and connection control functions. It is responsible for generating & managing signaling requests.

Management Plan:

Includes plane management, which performs management functions related to a system as a whole and provides coordination between all the planes, and layer management, which performs management functions relating to resources and parameters residing in its protocol entities.

5.3- ATM Devices and Interface:

An ATM network is made up of two types of general devices:

• ATM switch

• ATM endpoints

ATM switch:

ATM switches are responsible for cell transfer through an ATM network. In an ATM network ATM switches interconnected by point-to-point ATM links or interfaces.

ATM endpoint:

ATM endpoint contains an ATM network interface adapter.

ATM switches support two primary types of interfaces UNI and NNI. UNI connects ATM end systems to an ATM switch. It connects ATM switch to ATM endpoint devices. The NNI connects two ATM switches. These interfaces are further subdivided into private and public UNIs and NNIs.

Two types of User to Network Interfaces (UNI) first is Public UNI and Private UNI. And two types of Network to Network Interfaces (NNI) Private NNI (P-NNI) and Public NNI. Broadband Inter-Carrier Interface (B-ICI) connects two public ATM switches from different service providers.

5.4- ATM Cell:

ATM transfers data in fixed-size packet called cells. Each cell consists of 53 octets, or bytes.

5 bytes Header

48 bytes payload (user data)

Small, fixed-length cells are well suitable to transferring voice and video traffic at high data rates becausesuch traffic is intolerant of delays that result from having to wait for a large data packet to download, among other things.

The reason of fast processing in the ATM network, the ATM header has very limited function. Its main function is the identification of the virtual connection by an identifier which is selected at call set up and guarantees a proper routing of each packet. ATM allows an easy multiplexing technique of different virtual connections over a single link called virtual circuit. The Data field (48 bytes) length is very small. It decreases the internal buffers in the switches and queuing delays in those buffers are limited. Small buffers guarantee a small delay and a small delay jitter as required in real time systems. The data field of ATM cells is carried transparently through the network. No special processing is done on it inside the network. All services (voice, video, and data) can be transported via ATM, including connectionless services. ATM header contains following fields.

• GFC Generic Flow Control

• HEC Header Error Control

• CLP Cell Loss Priority

• VPI Virtual Path Identifier

• VCI Virtual Channel Identifier

• PT Payload Type

Generic Flow Control (GFC) Four bit field is used to control of cell flow only at the user-network interface

Virtual Path Identifier (VPI) in conjunction with the VCI identifies the next destination of a cell as it passes through a series of ATM switches on the way to its destination.

Virtual Channel Identifier (VCI) In conjunction with the VPI, identifies the next destination of a cell as it passes through a series of ATM switches on the way to its destination.

5.5- ATM Virtual Connections:

ATM networks are basically connection oriented. It established its connection virtually before it transfer data. Two types of ATM connections exist:

Virtual paths connections (VPC) identified by virtual path identifiers (VPI).

Virtual circuit connections (VCC) identified by the combination of a VPI and a virtual circuit identifier (VCI).

The idea behind dividing a virtual connection identifier into two parts is to allow hierarchical routing. Most switches in a typical ATM network do switching using VPIs. The switches at the boundaries of the network, those that interconnect directly with the endpoints, use both VPIs and VCIs. An ATM switch route the cell using both the VPIs and VCIs. A virtual path contains more then one virtual channels, all of which are switched transparently across the ATM network based on the common VPI. VPIs and VCIs have only local consequence across a particular link and are remapped, as appropriate, at each switch implemented with translation tables allows identifiers to be re-used. ATM switching can be done based on VP only (VP switching) or both VP and VC (VC switching).

5.6- Advantages of VP and VC:

The main advantage of Virtual Path (VP) and Virtual Channel (VC) is it simplifies network architecture. Network Transmission functions can be divided into those related to an individual logical connection (VC) and those related to a group of logical connections (VP). So on that bases it increased performance and reliability. It also reduced processing and short connection setup time. A large amount of the work is done, when VP is set up. New VC can be established by executing simple control functions at the end points of VP connection. No call processing is required at the transit nodes. The addition of new virtual circuits to an existing virtual path involves minimal processing. VP's simplified network architecture network transport functions can be divided into individual logical connection called virtual channel and a group of logical connections called virtual path. It also increased network performance and reliability. It reduced processing and short connection setup time. A large amount of the work is done when the virtual path is set up. By reserving capacity on a virtual path connection in expectancy of later call arrivals, new virtual channel connections can be established by executing simple control functions at the endpoints of the virtual path connection; no call processing is required at transit nodes. Thus, the addition of new virtual channels to an existing virtual path involves minimal processing. The virtual path is used internal to the network but is also visible to the end user. Thus, the user may define closed user groups or closed networks of virtual channel bundles.

5.7 ATM Layer Model:

In ATM reference model there are three layers .

• Physical Layer

• ATM Layer

• ATM Adaptation Layer

5.7.1- Physical Layer:

Physical layer is same as mentioned in OSI layer model. It manages the transmission dependent on medium type. ATM also supports wireless technology. Original design of ATM was based on SONET there are two reason to use SONET. One SONET provide high data rates and second reason is in SONET boundaries of cells are clearly defined.

ATM supports any technology of physical layer but the boundaries of the cell should be clearly defined. For the cell-based physical layer, no framing is forced. The interface structure consists of a continuous stream of 53-octet cells. Because there is no external frame imposed in the cell-based approach, some form of synchronization is needed. Synchronization is achieved on the basis of the HEC field in the cell header.

5.7.2 ATM Layer:

ATM layer combined with the ATM adaptation layer, the ATM layer is nearly equivalent to the data link layer of the OSI model. The ATM layer is responsible for cell multiplexing. Cell multiplexing means simultaneous sharing of virtual circuits over a physical link. It is also responsible for cell relay that means passing cells through the ATM network. To do this, it uses the VPI and VCI information in the header of each ATM cell. ATM layer provide routing, switching, multiplexing and traffic management. It is responsible for transmission and reception. Cell header generation and removal at source or destination is responsibility of ATM layer. Header format is different for UNI and NNI. ATM layer also responsible for following:

• Cell address translation

• Congestion Control/Buffer management

• Sequential delivery

• Provides routing, switching, multiplexing and traffic management.

• Transmission/Reception

• Cell header generation/removal at source/destination

• Header format is different for UNI and NNI

5.7.3- ATM Adaptation Layer (AAL):

The ATM Adaptation Layer (AAL) relays between ATM Layer and higher layer. When it receives information received from the higher layers it breaks messages to cells. AAL combined with the ATM layer nearly equivalent to the data link layer of the OSI model. TheATM Adaptation Layer (AAL)is responsible for dividing higher-layer protocols from the details of the ATM processes. And it converts higher layer user data for conversion into cells and segments the data into 48-byte cell payloads.

ATM Adaptation Layer is also responsible for some services. AAL is responsible for handling of transmission errors. The segmentation and reassembly it enables large blocks of data to be carried in payload field of ATM cells. It is also responsible for flow control. Several AAL protocols have been specified to meet the needs of different types of application.

ATM Adaptation Layer is subdivided into two logical layers:

• Convergence Sub layer (CS)

• Segmentation and Reassembly Sub layer (SAR)

5.7.3.1- Convergence Sub layer (CS):

In convergence Sub layer (CS) upper-layer frames are basic data units. It concerned with flow and error control for connection-oriented traffic

5.7.3.2- Segmentation and Reassembly Sub layer (SAR):

Segmentation and Reassembly Sub layer (SAR) is segments of upper-layer frames are basic data units. It concerned with segmenting frames at the source and reconstructing frames at the destination.

5.8- ATM Class of Services:

A class of Service is a type of traffic. For example the real time interactive class of service defines live video traffic. It is one of the most demanding applications with respect to data rates and network throughputs. Such traffic requires considerable amount of bandwidth.

An ATM network is designed to be able to transfer many different types of traffic at the same time, including real-time flows which required maximum bandwidth such as voice, video. With ATM, for very VCC that is set up, the customer specifies a desired class of service. Although each such traffic flow is handled as a stream of 53-octet cells traveling through a virtual channel, the way in which each data flow is handled within the network depends on the characteristics of the traffic flow and the requirements of the application. For example, real-time video traffic must be delivered within minimum variation in delay. The desired class of service determines the price for using a particular VCC.

ATM has defined 4 classes of service

CBR Constant Bit Rate

VBR Variable Bit Rate

ABR Available Bit Rate

UBR Unspecified Bit Rate

6- Multiprotocol Label switching:

Multiprotocol Label Switching (MPLS) is a multiprotocol technology of packet switching in connection oriented way. It called multiprotocol because it works with different protocols like Internet Protocol (IP), Asynchronous Transport Mode (ATM), and frame relay network protocols. IP is internetworking protocol which is global, but some other technologies are also very famous ATM is one of them. The relationship of ATM and IP is MPLS based on Tag switching. The concept of Tag Switching is given by CISCO when IETF start working on it use the term Label Switching.

Figurea-a shows, MPLS works between layer 2 and layer 3 in OSI layer model. MPLS is also called hybrid protocol because it works with different protocols. MPLS is a data carrying mechanism and it is connection oriented. MPLS connection oriented technology of packet switching networks.

6.1- MPLS Characteristic:

MPLS is a mechanism to manage traffic flows of with different flow management technique. It works in a different layer called layer 2.5 which is independent of Layer 2 and Layer 3. It maps IP addresses to fixed length labels but it uses the interfaces to existing routing protocols. It also supports ATM, Frame-Relay. In a traditional connectionless network, every router along the path makes an independent forwarding decision for every packet. In MPLS routing takes place only in the LER (Label Edge Router) routers. There are two types of router in MPLS, LER Label Edge Router and LSR Label Switched Router. Only LER are responsible for routing. LSP Label Switched router only in forwarding switches. Label switching increases speed and also provides Quality of Service (QoS). It hides the link layer and the differences between different Layer 2 protocols.

An only disadvantage of MPLS is we have to add an additional layer and the router has to understand MPLS.

6.2- MPLS Shim Header:

In MPLS with IP environment MPLS label is between layer 2 header and layer 3 headers as shim header. MPLS label is part of that shim header. The total size of MPLS shim header in IP networks is 32 bits. Which have four fields Label field, experimental field, BS flag, and time to live field.

• Label 20 bits

• Exp 3 bits

• BS 1 bit

• TTL 8 bits

If MPLS is running with ATM then it uses ATM Virtual Path Identifier (VPI) and Virtual Circuit Identifier (VCI) for his label. It uses VPI/VCI to write the label in ATM networks. In Frame relay environment it use Data Link Connection Identifier (DLCI) field.

6.3- MPLS Traffic Engineering:

MPLS require two types of routers LER Label Edge Router and LSP Label Switching Router.

6.3.1- Label Edge Router (LER)

Label Edge Router (LER) resides at the edge of an MPLS network and it is responsible to assigns and removes the labels from the packets.

6.3.2- Label Switching Router (LSR)

Label Switching Router is a high speed router in the MPLS network. ATM switches can be used as LSR without changing anything software or hardware because Label switching is equivalent to VP/VC switching. All the data transmission occurs on LSPs (Label Switched Paths). LER use routing and signalling protocols and establish a label switched path (LSP). The labels are distributed using Label Distribution Protocol (LDP) or Reservation Protocol (RSVP) or BGP (Border Gateway Protocol) and OSPF (Open shortest Path First). Each data packet encapsulates and carries the labels during its journey from source to destination.

The labels are used to forward the packets on LSP. The labels change during transmission on each node. Hardware switching is very quick packet switching technique.Label switching ingress devices use information in the layer-3 header to assign the packets to one of the predetermined paths - label assignment. The label accompanies the packet as it traverses the network. Subsequent routers along the path use the label to determine the next hop device. Label Switching devices only manipulate information in the label, processor-intensive analysis and classification of the layer-3 header occurs only at the ingress point.[25]

In figurea-a a MPLS network model is showing some Label Switched Paths (LSP). The network consists of LER and LSR routers. And they are showing three LSPs first is ALPSC from LER A to LER c. And Second label switched path is AMQTD which is from LER A to LER D. And last path is BMPSC. Now we will check the router table entries for these nodes all to gather and one bye one which will clear us how MPLS works.

For ALPSC

For AMQTD

For BMPSC

We can merge ASPSC & BMPSC

Sign up and be the first to receive our latest offers: