Destination and Next Hop Addresses

What is the relationship between the destination address in a datagram header and the address of the next hop to which the datagram is forwarded?The Destination IP address field in a datagram contains the address of the Ultimate destination.When a router receives a datagram,the router extracts the destination address D,and uses it to compute the address of the next router to which datagram should be sent,N.Although the datagram is sent directly to address N,the header in the datagram retains destination address D.All routes are computed using IP address.After computing the address of a next hop,N,IP software uses the address binding to translate N to an equivalent hardware address for transmission.

Physical Network Connection With Routers

The basic hardware component used to connect heterogeneous networks is a Router.Physically,routers resembles bridges that means each router is a special purpose hardware system dedicated to the task of interconnecting networks.Like a bridge,a contains a processor and memory as well as a separate I/O interface for each network to which it connects.The network treats a connection to a router, the same as the connection to any other computer.

The most important that you have to kept in your mind about router is that it is not only used to connect two similar technology,it can also be used to connect two different types of network technology. Router can connect two LAN's,a LAN and a WAN,or two WAN's.Furthermore,when a router connects two network in the same general category,the networks do not need to use the same technology.For ex:- A router can connect an Ethernet LAN to a frame relay network.Thus,each cloud represents an arbitrary network technology.On summarizing this whole,we can say that a router is a special purpose system dedicated to the task of interconnecting networks.A router can interconnect networks that use different technologies,including different media,physical addressing schemes or frame formats.

You can also visit my other blogs:
Operating System Concepts
Technomedia
Enter into world of Technology

Goal of Internet

The main goal of internetworking is to provide a universal service across heterogeneous networks.To provide universal service among all computers on an internet,routers must agree to forward information from a source on one network to the specified destination on the other network.The task is complex because as all we know that different frame formats and different addressing schemes are used by different networks.Hence,to overcome this problem protocol software is needed on computers and routers to make universal service possible.The main task of these internet protocols is to overcome the differences in frame formats and physical addresses to make communication possible among networks that use different technologies.
For knowing more about protocols for internetworking you can visit:
http://networkingconcepts4u.blogspot.com/2009/02/protocol-for-internetworking.html

Internet Architecture

With the help of routers it is possible for an organization to choose network technologies appropriate for each need,and to use routers to connect all networks into a single network.As all we know that router is used exactly two connections but commercial routers can connect more than two networks.Thus,a single router could connect any number of networks.However,an organization seldom uses single router to connect all of its networks.There are two main reasons behind it:

1.As all we know that router must forwarded each packet,the processor in given router is insufficient to handle the traffic passing among an arbitrary number of networks.

2.The second reason behind this is redundancy,because redundancy improves reliability.To avoid a single point of failure,protocol software continuously monitors internet connection,and instructs the routers to send traffic long alternative paths when a network or router fails.

Thus,when planning an internet,an organization must choose a design that meet's the organization's need for reliability,capacity and cost.In particular,the exact details of internet topology often depends on the bandwidth of the physical networks,the expected traffic,the organization's reliability requirements,and the cost and performance of available router hardware.

Connection between a NIC and a network

The type of connection used between a NIC and a network depends on that network technology.In some technologies,the NIC contains most of the necessary hardware,and attaches directly to the network medium using a single cable or optical fiber.In many other technologies,the NIC does not contain all the electronic circuitry needed to attach directly to the network.Instead,the cable from a NIC attaches to an additional electronic component that then attaches to the network.Surprisingly,the exact details of the connection between a NIC and a network are not determined by the technology that means a given network technology can support multiple wiring schemes.Let take the example of Ethernet technology in which technology i.e Ethernet remains same while the wiring scheme changes.Wiring scheme can be Thick Ethernet Wiring,Thin Ethernet Wiring and Twisted pair Ethernet Wiring.

Thin Ethernet Wiring

A second generation of Ethernet wiring uses a thinner,more flexible cable than the original thick wiring.Formally named as a 10 Base2 and informally named as thin wire Ethernet or Thinnet.This wiring scheme differs from Thicknet wiring scheme in three important ways.First,Thinnet generally costs less to install and operate than Thicknet.Second,because the hardware that performs the transceiver function is built into the NIC,no external transceivers are needed.Third,Thinnet does not use an AUI cable to attach the NIC to a communication medium.Instead,Thinnet attachs directly to the back of each computer using BNC connector.In a Thinnet installation,a coaxial cable stretches between each pair of machines.The cable doesn't need to follow a straight line-it may lie loosely on the tabletop between computers,run under floor,or run in a conduit.Although,wiring for a thin Ethernet appears to be completely different than the wiring for a thick Ethernet,the two schemes share several important properties.Both thick and thin cables are coaxial,meaning that they shield signals from outer interference.Both thick and thin cables requires termination,and both use the bus topology.Most important,because two wiring systems have similar electrical characteristics,that means signals propagate along the cable is in the same way.

Twisted Pair Ethernet

This type of Ethernet wiring is known as Third Generation of Ethernet wiring.This generation is completely differ from its previous two generation that are thick and thin Ethernet because it does not use a coaxial cable.This generation Ethernet typically known as twisted pair Ethernet or TP Ethernet and its first version was formally known as 10BaseT.The wiring for Fast Ethernet is known as 100BaseT,and wiring for Gigabit Ethernet is known as 1000BaseT.Instead of shared medium,the new technology extends the idea used with connection multiplexing in which an electronic device serves as the center of the network.
The basic electronic device is known as an Ethernet Hub.Like the other wiring schemes,twisted pair requires each computer to have a network interface card and a direct connection from NIC to the network.The connection uses unshielded twisted pair wiring with RJ-45 connectors which are the largest versions of the modular connectors used with the telephones.The connector on one end of a twisted pair cable plugs into the network interface on a computer and the connector on the other end plugs into the hub.
Thus,each computer has a dedicated connection to the hub device.Electronic components in hub emulate a physical cable,making the entire system operate like a conventional Ethernet.For ex:A computer attached to a hub must have a physical Ethernet Address that means each computer must use the CSMA/CD to access the network and the standard Ethernet Frame Format.In fact,software does not distinguish between thick Ethernet,thin Ethernet and twisted pair Ethernet because the network interface handles the details and hides any differences.

The Berkeley Broadcast Address Form

The University of California at Berkeley developed and distributed an early implementation of TCP/IP protocols as a part of BSD UNIX.The BSD implementation contained a non-standard feature that has affected many subsequent implementations.Instead of using a host suffix of all ones to represent a directed broadcast address,the Berkeley implementation used a host suffix that contained all zeros.The address form is informally known as Berkeley Broadcast.Unfortunately,many computer manufacturers derived their early TCP/IP software from the Berkeley implementation,and a few sites still use Berkeley broadcast.Some implementations include a configuration parameter that can select between the TCP/IPTCP/IP standard and Berkeley form;many implementations are built to accept both standard and Berkeley broadcast address forms.

Directed Broadcast Address

Sometimes,it is convenient to send a copy of a packet to all hosts on a physical network.To make broadcast easy,IP defines a directed broadcast address for each physical network.When a packet is sent to network's broadcast address,a single copy of the packet travels across the internet until it reaches the specified network.The packet is then delivered to all the hosts on the network.The directed broadcast address for a network is formed by adding a suffix that consists of all 1 bits to the network prefix.

To ensure that each network can have directed broadcast address,IP reserves the host address that contains all 1 bits.An administrator must not assign the all ones host address to a specific computer or the software may malfunction.If a network hardware supports broadcast,a directed broadcast will be delivered using the hardware broadcast capability.In such cases,transmission of the packet will reach all computers on the network.When a directed broadcast is sent to a network that does not have hardware support for broadcast,software must send a separate copy of the packet to each host on the network.

Authority for Addresses

Throughout an internet,each network prefix must be unique.For networks connected to the global internet,an organization obtains network numbers from the communication company that supplies internet connections.Such companies are called Internet Service Providers(ISPs).Internet Service Providers coordinate with a central organization,the Internet Assigned Number Authority,to ensure that each network prefix is unique throughout the entire Internet.For a private internet,the choice of the network prefix can be made by organization.To ensure that each prefix is unique, group that builds a private internet must decide how to coordinate network number assignments.Often,a single network administrator assign prefixes to all networks in the company's internet to ensure that numbers are not duplicated.

Loopback Address

IP defines a loopback address used to test network applications.Programmers often use loopback testing for preliminary debugging after a network application has been created.To perform a loopback test,a programmer must have two application programs that are intended to communicate across a network.Each application includes the code needed to interact with TCP/IP protocol software.Instead of executing each program on a separate computer,the programmer runs both program on a single computer and instructs them to use a loopback IP address when communicating.

When one application sends data to another,data travels from the protocol stack to the IP software,which forwards it back up through the protocol stack to the second program.Thus,the programmer can test the program logic quickly without needing two computers and without sending packets across a network.IP reserves the network prefix 127/8 for use with loopback.The host address used with 127 is irrelevant that means all host addresses are treated the same.By convention.programmers often use host number 1,making 127.0.0.1 the most popular form of loopback.During loopback testing no packets ever leave a computer that means the IP software forwards packets from one application program to another.Consequently,the loopback address never appears in packet traveling across a network.

Subnet and Classless Addressing

As the Internet grew,the original classful addressing scheme became a limitation.On one hand,the IP address space was being exhausted.On the other hand,because all networks had to choose one of the three possible sizes,that means many addresses were unused.Two new mechanisms were invented to overcome the limitations,known as subnet addressing and classless addressing.These two addressing schemes are so closely related that they can be considered to be a part of single abstraction.The generalization is straight forward that means instead of having three distinct address classes,allow the division between prefix and suffix to occur on an arbitrary bit boundary.Let understand this situation with the help of an example,Consider a network that contains 9 hosts.Only four bits of host suffix are needed to represent all possible host values.However,a class C address,which has the fewest address hosts possible,devote eight bits to the host suffix.Classless addressing solves the problem by allowing an ISP(Internet Service Provider) to assign a prefix that is 28 bits long that means network can have up to 14 hosts.

Protocol for Internetworking

As all we know that may protocols have been adapted for use in an internet,one suite stnds out as the most widely used for internetworking.The suite is formally known as The TCP/IP Internet Protocols.TCP/IP was the first set of protocols develop for use in an internet.Indeed,researchers who devised TCP/IP also developed the internet architecture.Work on TCP/IP began in 1970s,approximately the same time that Local Area Networks were being developed.The U.S military funded much of the research on the TCP/IP and internetworking through the Advanced Research Projects Agency.The military was among the first organizations to have multiple physical networks.Cosequently,the military was among the first to realize the need for universal service.By the mid-1980s,the National Science Foundation and the other U.S government agencies were funding development of TCP/IP and a large internet that was used to test the protocols.

Art of Protocol Design

Although the techniques needed to solve specific problems are well known,protocol design is on-trivial for two reasons.First to make communication efficient,details must be chosen carefully that means small design errors can result in incorrect operation,unnecessary packets or delays.For example:sequence numbers are usually stored in a fixed field in a packet header.The field must be large enough so sequence numbers are not reused frequently,but small enough to avoid wasting unnecessary bandwidth.Similarly,a higher level protocol can cause excessive overhead if the protocol chooses a non-optimal message size.

Second,protocol mechanism can interact in unexpected ways.For example:Consider the interaction between flow control and congestion control mechanism.A sliding window scheme aggressively use more of the underlying network bandwidth to improve throughput.A congestion control mechanism does,the opposite by reducing the number of packets being inserted to prevent the networking from collapsing.

The balance between sliding window and congestion control can be tricky,and a design that does both well is difficult.A protocol that is too aggressive about using bandwidth can adjust the underlying network and a protocol that is too conservative can result in lower throughput then necessary.More important,designs that attempts to switch from aggressive to conservative whenever congestion occurs tend to oscillate - then slowly increase their use of bandwidth until the network begins to experience congestion,decrease use until the network becomes stable and then begin to increase again.