News:

Choose a design and let our professionals help you build a successful website   - ITAcumens

Main Menu

Internet Routing

Started by VelMurugan, Aug 20, 2008, 03:18 PM

Previous topic - Next topic

VelMurugan

What is the Internet Routing

Can you imagine if every routers from its storage node can reach the point for each goal the information needed, it is possible that the router will accumulate a large routing table. As physical (cpu, memory) limitations on the router is simply not possible sometimes difficult to deal with a large routing table. Therefore, do not affect the ability to reach the destination in each case, we want to minimize the routing table. For example, a router connected to another router through a DS1 link connected to the Internet, then this can be a router nodes on the Internet all the information is stored, or it can also link all DS1, serial Of non-local information are stored. That is no router in its routing table in the storage of any data "package" to find the destination of non-local network of information, but these "package" sent to the other end of the serial link router From this router to provide the necessary information.

We often like this case we are talking about the serial DS1 link the other side of the router known as the "Gateway of Last Resort". This simple little trick for the routing tables can save 30 the number of entry-level. Routing information is not necessary to be too frequent exchange between the routers. Usually the routing tables in the blender to any router can provide the poor memory and CPU on a lot of unnecessary pressure. Copy the information should not affect the operation of the router to be transmitted. Although there is no need every millisecond refresh the routing tables are, of course, not every one week to refresh the routing tables. Routing is an important goal for the host to provide the ability to accurately reflect the current state of a network routing table.

Router is the most important operation will receive the packet sent to the correct path. Without routing the packets may lead to data loss. The routing table inconsistency will lead to routing loop and a packet in the interface between two adjacent cycle was sent. People very much hope that all the routers can have rapid convergence. Convergence can be defined as an informal measurement of all network routers unanimous view of the speed of the unit. It is hoped that the convergence is very small, because of this each router on the network even in the network topology (that is, network view) was a serious change in circumstances can accurately reflect the current network topology. When the network topology be changed, each router must transmit data to help other routers to the network convergence to the correct view. However, in refresh routing table when there is rapid convergence of its problems. If a link in the rapidly vibration (while disconnected, now closed), it will generate a lot of installation and withdrawal of the request. This link will eventually be exhausted every router on the network of resources, since other routers forced rapid installation or withdrawal of the routing. Therefore, even if rapid convergence is the goal of routing protocols, it is not all the network problems panacea.

Distance vector routing

Distance vector routing protocol routers to distribute all the neighbors to form a record list. These records for the network in each of the nodes of other non-node conferred on the overhead this value. It is worth noting that such information only distributed to the source router's neighbors router. Here's neighbors routers are often physically, but also in the eBGP apply to the logic of the situation. Overhead is the meaning of the source from the router to the target node link the sum of expenses. Source router regularly refresh its distance vector record and the record distributed to its neighbors router. O router will receive the record of the past and the present, if the past overhead smaller router along the vector from the past record of receiving the path referred to send output.

Many in the actual distance vector will be encountered when using the infinite problem. For example, we assume that all the links there is a cost for each unit and the adjacent node of the link between a corresponding unit. If the router to connect to the router X and Y Y router to connect to the router Z in Figure 1, we will find infinite problems. Y Z know to be an element of the overhead X and Z to know that two units of the overhead. YZ close link assumptions, the cost of this link has become infinite (Figure 2). Y Z now know that at the overhead is infinite, it will be this distance vector routing sent to X. Assumptions X Y then sent to a distance vector routing claim that it arrived at Z to two units of the overhead. Y now that it will be arrived at through X Z, it is sent to an updated version of the X distance vector routing at Z's claim that it is spending three units (Figure 3). Please note that X did not expect it to the Y this distance vector routing it is sent to the Y that distance vector routing to the projections. This is the distance vector routing the serious flaws in their not improve the structure of the barrier does not contain routing information. As shown in routers Legend will continue to change the path to Z of information. X and Y the two routers will never exchange the relevant Z router or the path information unit until the expenses reach a certain value of a prior agreement of the infinite value (for example, in RIP is 15)


X -------------------- Y -------------------- Z

Y: 1 X: 1 X: 2

Z: 2 Z: 1 Y: 1

[Figure 1]



X -------------------- Y -------- * *--------- Z

Y: 1 <------------- Z: infinite

Z: 2 -------------> X: 1

[Figure 2]



X -------------------- Y -------- * *--------- Z

Z: infinite (from Y) -> X: 1

Y: 1 <------------- Z: 3

[Figure 3]


The use of vector routing path can be resolved infinite problems. Each distance vector also including his adopted by the path (Figure 4). If the router to receive a path vector contains its own set records, the router will not set the record (Figure 5). Border Gateway Protocol (The Border Gateway Protocol) on the use of the above approach to resolving the problem of infinite. Obviously if you want to make the routing tables contain the transmission of the router AS (Autonomous Systems on the internet) path information, you will have to be inserted in the routing table to more information. Therefore, the designers decided to sacrifice BGP router that can withstand the effect of storage space and processing power.


X -------------------- Y -------------------- Z

Y: 1 (Y) X: 1 (X) X: 2 (YX)

Z: 2 (YZ) Z: 1 (Z) Y: 1 (Y)

[Figure 4]


X -------------------- Y -------- * *--------- Z

Y: 1 (Y) X: 1 (X)

Z: 2 (YZ) Z: infinity

[Figure 5]

Another infinite the settlement of the issue is the separation of the Road. Main idea is that if the adjacent routers in the path leading to the destination on the second node, adjacent to the router does not broadcast the router path. This solution can be used just examples to illustrate. Because the path to Z from the X through Y and then to Z, X and Y is because the neighbors router, the path from X, Y were not broadcast radio.



Link State Routing


A router using the link state routing, it will be on the network to all other routers distribute it to the router from its neighbors. This makes each router do not have to know from a source node to node purpose of the expenses, the routers can have a routing table. Loop the problem does not arise, because each router have the entire network topology. Main idea is a router emergence of three parts of the records contain source router (its own), o-o-router to router and costs.

Therefore, if the router through an A expenses for the three links connect to the router B, A and routers through a cost of 5 link to connect to the router C, then will be to the network router All of the router packet radio link state (LSPs) and. Each router will be able to receive the LSPs in a projected nodes to the purpose of the shortest path.

Obviously, LSP in the process of convergence is an integral part. If the network by adding the wrong LSP. Will lead to the wrong routing information (make longer than the original package along the path transmission) and even routing black hole. If the router C to the other routers to broadcast one of his neighbors router's path information, but the link disconnected router C withdrew earlier this broadcast. Unfortunately the first and second LSP first LSP first, then the other routers on the routing table does not correctly reflect the network topology, but can only wait until the right LSP to another.

To solve this problem, LSP introduced a sequence code. So all the routers on the network will be of some value as a Starting values to initialize their sequence code, and then broadcasting them in the LSP. This has solved the problem.

When using the sequence code sequence code will encounter this problem space is limited. LSPs can use the code sequence is set to have limited value. Therefore, when the serial code at the maximum, minimum start again from there. This router in comparison to the current link state records and set records difficult, because the code sequence of a priority. To solve this problem, for the LSP definition of one of the largest aging time. In other words, if the router X has not received within the time set records, it will discard the existing record left waiting for the updated records. Destinations attention to the need to make the path information is not valid. For example, when connected to a router Y, disconnect the network link, router to router Y Z broadcasting the message of this link, then the network routers are at this time Is that they can still reach the Z. If they are in the greatest time to receive less than an aging set records, they will assume that the link to the Y has not up. So that all the routers routing table will be the same, routers Y and Z can also use the limited sequence code.

The initialization code sequence is also the problem in another important respect. Y-router from the assumption that, at a time when networks started re-calculated path. When the router's link state began work agreement, it must know that it's re-initialization sequence code of values and other routers to make it consistent. Therefore, it broadcast a special with the initial set of path information. This record will tell the other routers that it needs to sequence code, and other routers will tell it.

Source : ArticleBase

sakunthala

his internet routing gives more explanation about the information