Original subtitles

What I'd like to point out once again is that each website is resolved to a different IP address.

Lastly, if I ping CNN dot com, notice that that also resolves to an IP address, but it's different

to the previous examples.

DNS is doing the name resolution, so it's resolving a domain name to an IP address.

And that's how I'm learning the IP address of CNN dot com or Google dot com.

You can ping many of the well-known websites on the Internet to find out what the IP addresses are.

You could also use any lookup which just does a DNS resolution of a domain name rather than trying to

ping the server.

So in summary, the devices on the Internet have been configured with IP version four addresses.

I'll explain more about the formatting of IP addresses in the next few minutes.

But for now, just take note that every device has an IP address and that includes my own machine.

The command IP config will show me the IP address on my local machine when using windows.

So in this example, my IP version four dress is 10 DOT 0.01 six.

You'll also notice here that I have an IP Version six address of 2001, Kolon 20, Kolon Kolon two.

In this video we're concentrating on IP Version four dresses, but in another video I'll explain IP

Version six, IP Version six is becoming more and more important because IP addresses are now exhausted

in certain parts of the world.

IP Version four or Internet Protocol Version four is a layer three or network layer protocol as per

the Odissi model.

In a different video, I explained the Osam model.

So if you're not sure about layers, please refer to that video.

IP Version four is a connection list protocol.

In other words, there are no sessions formed.

When traffic is transmitted, the transmitter simply sends data without notification to the receiver.

No status information is sent back from the receiver to the transmitter.

It's totally connection.

Less TCP, IP or transmission control protocol, on the other hand, is connection orientated.

TCP IP will set up a session so before transmission takes place in TCP IP, the transmitter sends what's

called a sun or synchronization message to the receiver.

There's a certain message from the receiver to the transmitter and then a EQ or acknowledgement message

from the transmitter to the receiver.

So before any data is transmitted, devices using TCP IP go through what's called the three way handshake.

Some Sinak and ACC IP, on the other hand, doesn't do any of that.

Each packet is treated independently of other packets.

That's why traffic can take different paths to get to a destination.

Rodders will route the traffic via different paths based on options such as load balancing because each

packet is independent and IP is a connectional as protocol.

Rodders can also base rodding decisions on different values such as bandwidth or hopp count.

But it is possible that packets from one session take divergent or different paths to get to a destination.

So, for example, Ruppe will base its routing decisions on Hopp count, which is not good, and hence

Ruppe is not used that often anymore.

OPF will base it on bandwidth.

Other routing protocols will use their own metrics to determine the best path.

I'll be discussing routing protocols in more detail later in this course.

But in brief, routing protocols determine the best path or best route from A to B..

This is based on the rockhill addressing structure in IP version for an IP version six, where we have

both a network and host portion.

As part of the address, routers base their routing decisions on the network portion of the address

rather than on the host portion of the address.

And I'll explain network and host portions in a moment.

IP also only gives best effort delivery of packets.

There is no guarantee of packet delivery.

Any packet could be misdirected, it could be duplicated or it could be lost in transmission when sent

to a destination.

And that should be expected in IP transmissions.

Once again, TCP IP, which is a connection orientated protocol, has the ability to transmit packets

that go missing UDP.

Another layer for protocol doesn't retransmit packets if they get dropped.

They simply lost and the applications need to take care of that.

There is also no data recovery features in IP.

If the packet, for example, gets corrupted, the end devices need to handle that and not the routers

in between.

So in summary, IP has no built-In sessions, no data recovery, no retransmissions.

Higher layer protocols such as TCP IP will need to handle dropped packets, corrupted packets, misdirected

packets and so forth.

IP does not provide those features and relies on higher layer protocols to implement those features.

So let's look at the format of an IP Version four address.

An IP version four dress is 32 bits in size, normally written in dotted decimal notation such as this

example 10.1 dot one that one each of value such as ten is eight bits in size.

So in other words, we have X, X, X, X with each X being eight bits in length, also known as an

octet.

The total size of the address is 32.

But please refer to the binary video if you're not sure about bits and how to convert this address into

binary and back again.

IP addresses once again have a hierarchical structure to enable routing, which consists of two main

parts.

We have the network portion of an address and the host portion, and we look at that in more detail

in a moment.

IP addresses are used for routing in a very similar way to the way DHL or FedEx route parcels.

Based on a destination address, rodders will route traffic to a destination address when unicast packets

are transmitted.

Multicast packets use a different mechanism and do the routing based on source address.

So as an analogy, DHL or FedEx, or sending the parcel to a destination based on the destination on

the Postle routers are sending packets to destinations based on the destination address in the packet.