ROUTING PROTOCOLS: A router constructs its routing table using the information it receives from other routers.

The router changes its routing table in response to routing updates that provide additional information or notification that conditions in the network have changed (for example, a link has failed). This responsiveness explains why using a routing protocol is often called dynamic routing. A routing protocol governs how routers exchange routes and other network information with each other. A routing protocol governs how routers exchange routes and other network information with each other. The protocol must dictate parameters such as the following: How routers compute a routes metric and select the best route for their routing table. What information routers include in the update message it sends to the neighboring routers When routers send and receive updates and hellos we can lower overhead and conserve bandwidth by altering how often routers send certain messages. EXISTING PROTOCOLS FOR INFRASTRUCTURELESS NETWORKS: 1) Routing Information Protocol(RIP): RIP is a well-known and commonly used distance-vector routing protocol. Although originally developed for LANs, RIP can also be used in WANs. RIP is simple to configure but can be slow to converge. Because route selection relies purely on hop count, RIP may not always generate the best routes for WANs, which usually include links of varying bandwidth. In such an environment, the lowest hop count is not always the fastest or best route. RIP uses the Bellman Ford Algorithm to calculate the network topology were each router sends list of distance-vectors each of its neighbors periodically. The metric used must be a positive integer. This metric measures the cost to get to the destination. In RIP, this cost describes number of hops. Here the distance-vector is a vector containing the following elements (destination, cost, source) where destination is the destination network, cost is the sending router's metric to the destination + 1 ,source is the sending router's id. One distance-vector is created for each entry in the routing table, and the entire set of distance-vectors are sent periodically.To avoid the problem of infinity, this protocol limits the diameter of the routing domain to hops.RIP v2 overcomes v1 disability of communicating other autonomous system. RIP Message Format: 1)Command: specifies the type of message: request (1) or response (2) 2)Version: 1 or 2 3)Family: TCP/IP family (2) 4) Network address: destination network address 5)Distance: the hop count from the advertising router to the destination network RIP uses three timers to support its operation Periodic timer which does the job of advertising of regular update messages, Expiration timer which governs the validity of a route. When the time is timeout, the hop count of the route is set to 16 and the Garbage collection timer which eliminates invalid routes as they are not immediately removed from the table until the garbage collection timer is timeout.

Therefore the disadvantage arises the routing information protocol can only use hop count as its metric. The other drawback is that during each routers updates it transmits the entire routing table.

2) Open Shortest Path First (OSPF): OSPF is a link-state routing protocol for Internet Protocol networks. It uses a link state routing algorithm and falls into the group of interior routing protocols which operate within a single autonomous system. OSPF was designed to cope with several of RIPs shortcomings. For example, OSPF provides quicker convergence and more sophisticated methods of computing best routes using Djikstras algorithm. Instead of sending routing table entries, routers send link state advertisements (LSAs) that allow peers to construct a more comprehensive, accurate, moment-to-moment topology of the network. In link state routing, if each node in the domain has the entire topology of the domain, the node can use Dijkstras algorithm to build a routing table Using OSPF, a host that obtains a change to a routing table or detects a change in the network immediately multicasts the information to all other hosts in the network so that all will have the same routing table information. Unlike the RIP in which the entire routing table is sent, the host using OSPF sends only the part that has changed. With RIP, the routing table is sent to a neighbor host every 30 seconds. OSPF multicasts the updated information only when a change has taken place. Rather than simply counting the number of hops, OSPF bases its path descriptions on "link states" that take into account additional network information. OSPF also lets the user assign cost metrics to a given host router so that some paths are given preference. OSPF supports a variable network subnet mask so that a network can be subdivided. An AS can be divided into a number of areas, which are groups of contiguous networks and attached hosts. Routers with multiple interfaces can participate in multiple areas. These routers, which are called Area Border Routers, maintain separate topological databases for each area. OSPF Message Format : Fixed Header: All OSPF packets begin with a 24-byte header. Version numberIdentifies the OSPF version used. TypeIdentifies the OSPF packet type as one of the following: HelloEstablishes and maintains neighbor relationships. Database descriptionDescribes the contents of the topological database. These messages are exchanged when an adjacency is initialized. Link-state requestRequests pieces of the topological database from neighbor routers. These messages are exchanged after a router discovers (by examining databasedescription packets) that parts of its topological database are outdated. Message lengthSpecifies the packet length, including the OSPF header, in bytes. Source Router IP addressIdentifies the source of the packet. Area IDIdentifies the area to which the packet belongs. All OSPF packets are associated with a single area. ChecksumChecks the entire packet contents for any damage suffered in transit. Authentication typeContains the authentication type. All OSPF protocol

exchanges are authenticated. The authentication type is configurable on per-area basis. AuthenticationContains authentication information. DataContains encapsulated upper-layer information Border Gateway Protocol (BGP): BGP is a standardized exterior gateway protocol (EGP), as opposed to RIP, OSPF, and EIGRP which are interior gateway protocols (IGPs). BGP Version 4 (BGPv4) is the current standard deployment. BGP is considered a Path Vector routing protocol. BGP was not built to route within an Autonomous System (AS), but rather to route between ASs. BGP maintains a separate routing table based on shortest AS Path and various other attributes, as opposed to IGP metrics like distance or cost. BGP is the routing protocol of choice on the Internet. Essentially, the Internet is a collection of interconnected Autonomous Systems. BGP Autonomous Systems are assigned an Autonomous System Number (ASN), which is a 16bit number ranging from 1 65535. A specific subset of this range, 64512 65535, has been reserved for private (or internal) use. BGP utilizes TCP for reliable transfer of its packets, on port 179. BGP is not a necessity when multiple connections to the Internet are required. Fault tolerance or redundancy of outbound traffic can easily be handled by an IGP, such as OSPF or EIGRP. BGP is also completely unnecessary if there is only one connection to an external AS (such as the Internet). There are over 100,000 routes on the Internet, and interior routers should not be needlessly burdened. BGP should be used under the following circumstances: Multiple connections exist to external ASs (such as the Internet) via different providers. Multiple connections exist to external ASs through the same provider, but connect via a separate CO or routing policy. The existing routing equipment can handle the additional demands. BGPs true benefit is in controlling how traffic enters the local AS, rather than how traffic exits it.

Reliable Routing over a logical network:

Routing over a Binomial Graph:
A scalable and fault tolerant topology called binomial graph (BMG), provides desirable topological properties in terms of both scalability and fault-tolerance for high performance computing such as reasonable degree regular graph, low diameter, symmetric graph, low cost factor, low message traffic density, optimal connectivity, low fault-diameter and strongly resilient. Several fault-tolerant routing algorithms are provided on BMG for various message types. More importantly, BMG is able to deliver broadcast messages from any node within log2(n) steps.

A circulant graph with n nodes and jumps j1, j2, ..., jm is a graph in which each node i , 0 <= i<=n1, is adjacent to all the vertices ijk mod n, where 1 <=k<=m. A binomial graph

network (BMG) is a circulant graph where jk is the power of 2 that is less thanor equal to n. g highly dynamic applications.Binomial graph (BMG) provides desirable topological properties in terms of both scalability and fault-tolerance for high performance computing such as reasonable degree, regular graph (every node has the same degree), low diameter, symmetric graph (in the sensethat an average inter-nodal distance is the same from any source node) and low cost factor. It also has low message traffic density, optimal connectivity, low fault-diameter, is strongly resilient and has good optimal probability in failure cases.BMG is an undirected graph G:=(V,E) where V is a set of nodes (vertices); |V | = n; and E is a set of links (edges). Each node i, where i2V and i=0,1,...,n-1, has links to a set of nodes U, where U={i1,i2,...,2k|2k _ n} in circular space, i.e., node i has links to a set of clockwise (CW) nodes {(i+1) mod n, (i+2) mod n,..., (i+2k) mod n | 2k <=n} and a set of counterclockwise (CCW) nodes {(n+i -1) mod n, (n+i-2) mod n,..., (n+i-2k) mod n | 2k <=n}. The structure of BMG can also be classified in the Circulant graph family1. A Circulant graph with n nodes and jumps j1, j2, ..., jm, where m 2 N, is a graph in which each node I , 0 <=i <= n 1, is adjacent to all the vertices i jk mod n, where 1<=k<=m. Original routing algorithms This section presents three original two-terminal routing algorithms . One optimal routing algorithm is based on breadth-first search, and the two suboptimal routing algorithms are called basic and variant. Each node in the graph may run the same routing algorithm because all nodes in BMG are equal (both regular and symmetric). 2.1 Breadth First Search Optimal Algorithm The optimal routing algorithm can use a breadth first search technique with a modified graph coloring algorithm. Although this algorithm gives the optimal result, the complexity of the algorithm is O(Degree).

Basic Sub Optimal Algorithm A basic algorithm to estimate the shortest path between nodes is to use a rule-based method that sends the unicast messages to a neighbor that has the closest ID to the destination ID as shown in Algorithm 2.The complexity of the basic unicast routing algorithm is O(Degree). Variant Sub Optimal Algorithm This algorithm is the variant of the basic algorithm that allows messages to go forward to a neighbor of which ID is not the closest ID to the destination ID if the destination is directly connected to the neighbor.The complexity of the variant unicast routing algorithm is second order of Degree. New routing algorithms:

In order to always stay on the shortest path from a source to a destination, messages must be delivered through a neighbor that has the estimated shortest hop to the destination unlike the original basic sub-optimal algorithm that estimates the shortest distance between neighbors and destination (i.e., a greedy algorithm). The key to success of this algorithm is how well we can estimate the number of hops that is used for sending messages between two nodes. Bit Counting method The bit counting method represents the distance between a source and a destination in a binary format.The bit-1 represents the number of hops that messages can travel, e.g., if a distance between a source and a destination (|destIDsrcID) is 9 (binary is 1001), a message is forwarded to nodes with distance 8 (1000) and 1 (0001). Hence the message can be delivered within two hops. From the above example, it does not matter which of the distances is selected as the first hop.Thus, load balancing of both links and neighbors can be implemented by a node if the next neighbor is randomly selected from all those candidates. Quality of service (QOS) can also be implemented by a node simply by selecting the next hop based on the priority of its candidate neighbors.The bit counting method can be used to estimate the number of hops by counting the number of bit-1 of distance between the source and the destination in both clockwise and counter-clockwise directions. The estimated number of hops is the minimum number of bit of both directions. Consecutive Bit Elimination Method Estimating the number of hops using the bit counting method might be too pessimistic, e.g., if a distance between a neighbor and a destination (|destIDneighborID|) is 7 (binary is 0111), the bit counting method will estimate the number of hops is 3, jumping to nodes with distance 4 (0100), 2 (0010) and 1 (0001),respectively. However, by using the counter-clockwise links we can use only 2 hops by jumping to distance 8(1000) on one direction and 1 (0001) on the other direction (7 = 8-1). Hence we may estimate more precisely the number of hops between two nodes by eliminating the sets of consecutive bits before counting the number of bits in both directions. Consecutive bits elimination can be performed by adding the value of the least significant consecutive bit to the distance between source and destination in both clockwise (CW) and counter clockwise (CCW) directions. This procedure is repeatedly performed until there is no consecutive bit left or the result after adding the values is more than jm, where jm is a maximum power of 2 that is less than or equal to n.For example, a distance in clockwise direction is 110 (the binary is 1101110), i.e. CW = 110 and CCW = 0.Consider the distance binary 1101110, the first group of consecutive bits from the right is 1110; therefore, the least significant, consecutive bit is 2 (binary is 10). The first step is performed by adding 2 (10) to both CW and CCW directions. Hence, the distance 110 can be routed by jumping with distance 112 (binary is 1110000) in CW and 2 (binary is 10) in CCW. Notice that the bi nary of 112 (1110000) still has a consecutive bit, thus the next value to add in both directions is 16 (binary is 10000). After adding 16, distance 110 can be jumped with distance 128 (biary is 10000000) in CW and 18 (binary is 10010) in CCW. In conclusion, the routing for a distance of 110 in the clockwise direction can be performed within three hops, i.e., one jump in a clockwise direction with the distance 128 and two jumps in a counter-clockwise direction with the distance 2 respectively 16. Again, it does not matter the order in which these jumps are undertaken.

The complexity of this algorithm is O(log(n)). How do we detect Faults? Self-healing capability of BMG This section presents the self-healing capabilities of BMG as a solution to prevent potential problems of network bisection and a decrease in routing performance when the network has a high percentage of failed nodes. Self-Healing Methods There are two methods presented in this section. The first approach is called the naive method. This method destroys the original network and reconstructs the BMG with the remaining nodes. The second method is called adaptive method. It only destroys and reconstructs the links that are different between the original BMG and the BMG after excluding of all failed nodes. Naive Method: This is the simplest method to reconstruct the BMG topology. Suppose there are F nodes in BMG size N. There are two steps involved in this method. The first step removes all existing links. The second step establishes all connections of BMG size N F. For each link in the BMG, a node that has a higher ID will initiate the connection to the node that has a lower ID.