BASIC ROUTER

BASIC ROUTER

Before we learn more about how to configure cisco router, we need to understand better about some basic rules of routing. Also of course we must understand the numbering system IP, subnetting, netmasking and his brothers.

Case in point:

Host X à 128.1.1.1 (ip Class B network id 128.1.xx)

Host Y à 128.1.1.7 (class B IP network id 128.1.xx)

Host Z à 128.2.2.1 (class B IP network id 128.2.xx)


In the case above, the host of X and Y hosts can communicate directly but both hosts X and Y can not communicate with host Z, because they have different network Id. How to communicate with Z so that X and Y? use a router!

Example of use subnetting

Host P à 128.1.208.1 subnet mask 255.255.240.0

Host Q à 128.1.208.2 subnet mask 255.255.240.0

Host R À 128.1.80.3 subnet mask 255.255.240.0

Well, when subnetting is used, then the two hosts are connected to the same network segment can communicate only if both the network id and his subnetid sesuai.Pada case above, P and Q can communicate with directly, R has the same network id P and Q but has a different subnetidyang. Thus R can not communicate directly with P and Q. How can R be able to communicate with P and Q? use a router!

So the function of the router, it is easy to say, connecting two different networks, precisely directs the best route to achieve the expected network

In its implementation, the router is often used to connect a network between institutions or companies, each of which has a network with different network id. Another example that is currently popular is when your company will be connected to the internet. Then the router will function data packet stream from your company to other institutions via the Internet, of course, your network number would calm with perushaaan that you want to go.

If you simply connect the 2 pieces of tissue, in fact you can also use the pc-based Windows NT or Linux. By providing 2 pieces of network cards and some settings, you actually have to make practical router. But of course with all faults.

In the market is very diverse brand routers, among others baynetworks, 3com and Cisco. The course modules we will discuss specific cisco. Why? because Cisco is a router that is widely used and much used as a standard for other products.

More info on routing

The data from the devices that connect to the Internet are sent in the form of datagrams, ie data packets defined by the IP. Datagram has a destination address of data packets; Internet Protocol check this address to deliver a datagram from the device of origin to destination devices. If the datagram destination address is located one network with the device of origin, the datagram directly delivered to the destination device. If you find that the datagram destination address does not exist in the same network, datagram delivered to the most appropriate router (the best available router).

IP Router (commonly called routers only) is a device that performs the function of an IP datagram forwarding on the network layer. The router has more than one antamuka network (network interface) and can forward a datagram from one interface to another interface. For each datagram is received, the router checks whether the datagram is addressed to him. If it is not addressed to the router, the datagram is delivered to the transport layer.

If the datagram is addressed to the router, which will be examined is that its forwarding table to decide where the datagram should be addressed. Forwarding table is a table consisting of pairs IP address (host address or network address), following the router address, and the interface where the outgoing datagram.

If you do not find a line that was in the forwarding table according to the destination address, the router will give a message to the sender that the address in question can not be achieved. This incident can be analogous with the message "return to sender" on the regular mail. A router can also tell that he was not the best router to a destination, and suggested the use of other routers. With these three functions contained on this router, the hosts on the Internet can connect to each other.

Static and Dynamic

In general, routing coordination mechanisms can be divided into two: static routing and dynamic routing. In static routing, the entries in the router's forwarding table is filled and removed manually, while the dynamic routing changes made through routing protocols. Routing Static routing is the simplest arrangement that can be done on a computer network. Using pure static routing in a network means to fill each entry in the forwarding table at each router in the network.

The use of static routing in a small network of course is not a problem, only a few entries that need to be loaded on the forwarding table at each router. But you certainly can imagine what if must complete the forwarding table at each router that amount is not small in a large network. Especially if you are assigned to fill in the entries on all routers on the Internet which in large amounts, and growing every day. Sure bother at all!

Dynamic routing is a means used to remove the obligation to fill the forwarding table entries manually. Routing protocol to set the routers that can communicate with each other and give each other routing information that can change the contents of the forwarding table, depending on the state of its network. In this way, the routers know the final state of the network and capable of forwarding the datagram in the right direction.

Interior Routing Protocol

In the early 1980s, the Internet is limited to the ARPANET, Satnet (expansion of the ARPANET which uses satellite), and several local networks that connect through the gateway. In its development, the Internet requires a hierarchical structure in anticipation of a network that has to be big. Internet and then broken up into several autonomous system (AS) and currently consists of thousands of U.S. Internet. Every American has a mechanism for exchanging and collecting its own routing information.

A protocol used to exchange routing information within the United States classified as an interior routing protocol (IRP). Results of collecting routing information is then submitted to the U.S. in the form of reachability information. Reachability information issued by a United States contains information about networks that can be achieved through the United States and become U.S. terhubungnya indicator to the Internet. Submission of inter-AS reachability information carried out using a protocol that is classified as an exterior routing protocol (ERP).

IRP is used as standard on the Internet to date is the Routing Information Protocol (RIP) and Open Shortest Path First (OSPF). In addition to these two protocols have also routing protocols that are proprietary but widely used on the Internet, namely Internet Gateway Routing Protocol (IGRP) from Cisco Systems. IGRP Protocol later expanded into the Extended IGRP (EIGRP). All of the above routing protocols use metrics as a basis for determining the best path that can be taken by the datagram. Metrics associated with "costs" contained in each link, which can be either throughput (data rate), delay, connection fees, and reliability of the link.

I. Routing Information Protocol

RIP (acronym, pronounced as rip) included in the distance-vector protocol, a protocol is very simple. Distance-vector protocol is often also called Bellman-Ford protocol, because it comes from the shortest distance calculation algorithm by RE Bellman, and is described in the form of the algorithm was first distributed by Ford and Fulkerson.

Each router with distance-vector protocol when it was first run only know how to route to itself (local information) and do not know where he is the network topology. The router then sends the local information in the form of distance-vector to all the links that connect directly with it. Router that receives the information to calculate distance-vector routing, adding distance-vector with a link metric information received, and put it into the forwarding table entry if it deems is the best path. After the addition of metric routing information is then sent back to the router interface, and is performed every certain time interval. So forth so that all routers in the network knows the network topology.

Distance-vector protocols have weaknesses which can be seen if there is a link in the network are disconnected. Two possible failures that may occur is a bounce effect and the count-until-no-till (counting to infinity). Bounce effects can occur in networks that use different metrics at least a link. Links are broken can lead to routing loops, so the datagram that passes through a particular link only spinning between two routers (bounce) until the age (time to live) datagram is over.

Counting-to-no-till occurs because the router was too late to inform you that a link is lost. This delay causes the router to send and receive distance-vector and calculate the metric up to a maximum distance-vector metrics achieved. The link is declared off after distance-vector reached the maximum metric. When calculating this metric also occur routing loops, even for a longer time than in case of bounce effect ..

RIP did not adopt the distance-vector protocol for granted, but by doing some additions to the algorithm for routing loops that occur can be minimized. RIP Split horizon is used to minimize the bounce effect. Split horizon principle used is simple: if node A submits a datagram to the destination X through node B, then for B does not make sense to achieve the goal of X by A. So, A does not need to tell B that X can be accomplished B through A.

To prevent cases of count-until-no-till, RIP Triggered Update method. RIP has a timer to know when the router must re-routing information. If there is a change in the network, while the timer is not expired, the router still need to send routing information because it was triggered by the changes (triggered updates). Thus, the routers in the network can quickly find out the changes and minimize the possibility of routing loops occur.

RIP is defined in RFC-1058 using the metric between 1 and 15, while 16 are considered as no-till. Route with a distance-vector 16 is not included in the forwarding table. This 16 metric limits prevent the count-until-no-till is too long. RIP packets are normally sent every 30 seconds or sooner if there is triggered updates. If within 180 seconds a route is not updated, the router route delete entries from the forwarding table. RIP does not have any information about the subnet route. The router should treat each route has received the same subnet with the subnet on the router. Thus, RIP does not support Variable Length Subnet Masking (VLSM).

RIP version 2 (RIP-2 or RIPv2) attempt to generate some improvements over RIP, namely support for VLSM, authentication, provide the following information hop (next hop), and multicast. Addition information subnet mask on each route to make the router do not have to assume that these routes have the same subnet mask with the subnet mask is used to it.

RIP-2 also use authentication for routing information to find out which ones can be trusted. Authentication is required on the routing protocol to create a more secure protocol. RIP-1 does not use authentication so people can give false routing information. Next hop information in RIP-2 is used by routers to inform a route but to reach the route does not pass through routers that provide information, but other routers. Use the following hops are usually on the border between the United States.

RIP-1 uses the broadcast address to send routing information. As a result, this packet is received by all hosts within the subnet and add the host's workload. RIP-2 can transmit IP packets using multicast on 224.0.0.9 so that not all hosts need to receive and process the routing information. Only routers that use RIP-2, which receives the routing information without the need to disrupt other hosts in the subnet.

RIP is a routing protocol that is simple, and this is the reason why RIP is the most widely implemented in the network. Set up routing using RIP is not complicated and gives quite acceptable results, especially if a network link failure rarely occurs. However, for large networks and complex, RIP may not be enough. In such circumstances, the RIP routing calculation often takes a long time, and lead to routing loops. For a network like this, most computer network specialists using a protocol that included in the link-state

II. Open Shortest Path First (OSPF)

Link-state technology was developed in the ARPAnet to produce a distributed protocol which is much better than distance-vector protocol. Instead of exchanging distance (distance) to the destination, each router in the network has a network map that can be updated quickly after any topology change. This map is used to calculate the route which is more accurate than using distance-vector protocol. The development of this technology eventually produce protocol Open Shortest Path First (OSPF) developed by the IETF for use on the Internet. Even now the Internet Architecture Board (IAB) has recommended OSPF instead of RIP.

The principle of link-state routing is very simple. In lieu of calculating route "best" by a distributed, all routers have a network map and count all the best routes from this map. Network map is stored in a database and each record in the database specifies a link in the network. Records are sent by routers that connect directly with each link.

Because each router need to have a network map that illustrates the last condition complete network topology, any changes in the network must be followed by changes in link-state databases are located at each router. Link status changes are detected the router will change the database link-state router, then the router sends the changes to other routers.

A protocol used to transmit this change must be fast and reliable. This can be achieved by flooding protocol. In the flooding protocol, the message sent is a change from the database and serial number of the message. By only sending data base changes, the time required for delivery and message processing fewer dibandingdengan send the entire contents of the database. Message sequence number is needed to determine whether the message received is more recent than that contained in the database. This serial number handy in case of broken links become connected again.

At the moment there are broken links and a separate network, the second database into different parts of the network. When the link is broken back to life, the database on all routers must be equated. This database will not return the same as sending one message link-state only. The process of equating the data base on the neighboring router is called the turn adjacency. Two neighbors called the adjacent routers when link-state database has both the same. In this process the two routers are not exchanging data base because it will take a long time.

The process consists of two-turn adjacency fasa.Fasa first, the two routers exchange database description is a summary of the database that each router. Each router then compares the description of the basis of data received on the basis of available data. In the second phase, each router asks neighbors to send records of different databases, ie if the router has no such record, or the serial number of records that are owned less than that delivered by a description of the database. After this process, the router update multiple records and is then sent to other routers via flooding protocol.

Link-state protocol better than distance-vector protocol is caused by several things: the time required to converge faster, and more importantly, this protocol does not produce routing loops. This protocol supports the use of multiple metrics at once. Throughput, delay, cost, and reliability are the metrics that are commonly used in the network. In addition, this protocol can also generate a lot of paths to a destination. Suppose router A has two paths with equal metrics to host B. Protocol can enter both paths into the forwarding table so the router can share the load between the two routes.

Draft OSPF uses link-state protocol with some additional functions. These functions were added, among others, supports multi-access networks, such as X.25 and Ethernet, and divide large networks becoming some areas.

It has been described above that each router in link-state protocols necessary to establish neighbor adjacency with the router. On multi-access network, each router's neighbors to more than one. In these circumstances, every router in the network need to establish adjacency with all other routers, and this is inefficient. OSPF adjacency streamline this by introducing the concept of designated routers and backup designated routers. All routers need only adjacent to the designated router, so that only the designated router that is adjacent to all other routers. Designated backup router will take over the functions of a designated router that failed to function.

The first step in a multi-access network is to choose a designated router and back up. This selection is included in the Hello protocol, the OSPF protocol to find neighbors of a router in each link. After the election, then the routers to form adjacency with the designated router and back up. Any changes in the network, the router sends a message using a flooding protocol to the designated routers and designated router that sends the message to other routers in the link.

Designated backup routers also listen to the messages being sent to the designated router. If the designated router fails, the backup designated router then becomes the new and selected a new backup designated router. Due to the new router has been designated adjacent to the other routers, not necessary anymore equating process database that takes a long time is.

In a large network is certainly needed a large data base is also to save the network topology. This leads to the need for a larger router memory and route calculation time is longer. To anticipate this, OSPF uses the concept of area and backbone. The network is divided into several areas that are connected to the backbone. Each area is considered as a separate network and the routers in it only needs to have a network topology map in the area. Routers located on the border between the area just send a summary of the links contained in the area and do not send one area topology to other areas. Thus, the calculation becomes much simpler route.

Simplicity vs.. Ability

We already see at a glance how the RIP and OSPF work. Each routing protocol has its pros and cons of each. RIP protocol is very simple and easy to implement but can lead to routing loops. OSPF protocol is a protocol that is more complex and better than RIP but requires memory and CPU time are great.
In many places there are also using a combination of static routing, RIP, RIP-v2, and OSPF. The results in this network indicates that the administration of static routing is much more time consuming than dynamic routing. Observations on the dynamic routing protocol also indicates that the RIP uses much more bandwidth than OSPF and the bigger the network, bandwidth used RIP to grow larger as well. So, if you are designing a TCP / IP is certainly large OSPF routing protocol is the right choice.