DCN January 26. 2007
Trying to understand how Ethernet work? What is Ethernet?
Go back in history, the island of Hawaii in the Pacific, in the late 60’s the university of Hawaii had a very clever thought, they had some offices on one side of the island and on the other side, they wanted to communicate between these offices, and their was no way for them to do this, they invented a method, which was satellite service, they developed their own little protocol to send messages to each other, they sent the messages to the satellite and it was a source address and a destination address and of course data, if add wanted to send to add four, the messages would be sent to the satellite and rebound it to the computer. Lets say SA = 1 DA=4, Three would receive the message and ignore it and would only go to its destination. Primitive and it worked. This was called Aloha net. They thought to themselves to send a single down through copper wire. There was a group in the 60’s that developed things like mice, but the most important was Ethernet, and the most was Zerox. ThreeCOM manufacturing Ethernet cards, at the same time IBM was inventing token ring, yet today token ring is gone, Ethernet survived and Ethernet is thriving, December of last year Ethernet people put the standards at running Ethernet at 10 billion bits per second, Then the standardized of 100 billion bits per second. Macaffe invented Etherenet, they put in strands of coaxial cable, DataCom cables coaxial and twisted pair cables, Coax has a thick piece of copper surrounded by a insulated matieral (plastic), surround by a weave, amazing thing about coax the copper is thick and can transfer a lot of Hrz, (hrz dead German guy, working in the 18’s with electricity), the number of wave cycles per second is called one Hrz(hertez), a million Hrz means 100 hrz per second, coax cable b.c its nice and thick can theoretical can handle 100, mill hrz. The coax in the tv can handle 600 mil hrtz, Each station takes 600mil hrtz, Back to MacAfee, we invent a little plug that attaches in the coax, in every comp a special hard that will sit in the expansion slot in each computer (THE NIC) the Network Interface Card has logic, try in to visualize the NIC is a separate device, and has simple logic burned into it, and works with your computer, The idea about the cable was SPEED, A NIC using Ethernet would take data, Operating software uses the inturp and inturrpts the cpu and sends the data, INT52 is a well known interrupt and goes through a specific area in ram grabs the data and moves it to a NIC, The Ethernet guys MacAfee used the follow rules: First the Maximum Transition Unit (MTU), 100 protocols Each has a MTU, All protocols take data, and break it in to pieces the maximum transition unit size is going to be 1500bytes, if you wanted to send 1501 bytes, it would resolve in two transmissions, that is just Ethernet, remember this and remember it NEXT SEMISTER. As the same time in the Ethernet guys IBM was inventing token ring, TOKEN RING when they designed it could send 64,000bytes (huge packets of information), we are going to figure out why the groups made these bye decision Ethernet primitive simple protocol survived, and Token ring survived. When ever a int was sent the NIC would build a frame of data, and infront of it there is a Header, Ethernet header very simple while the TCPIP complicated, HDR has DA SA TL Destination Address, Source Address, A one byte fule called either TYPE or LENGTH. Ethernet uses MAC address, All LAN addresses is how many bytes 4 bytes ip and MAC address is 6 bytes, trillion of address, IMPORTANT’ the MAC address is made up of TWO components, NO two MAC address can be the same, Network standards, the IEEE group Institute of Electronic Engineering (A GROUP), OKAY you are company 999, the first three bytes of the 6 byte address is Manufacture ID, They are identical to AHOLA net, Workstation one wants to send data to Workstation three, the NIC would build the bytes and send it to machine two and ignore the address, and send it to three. Ethernet survived and why…Ethernet was simple that’s the answer it was so basically simple that it thrived, probably for the rest of your lives, you will be working with Ethernet. Test 1 DCN386 questions TWO How does Ethernet work…here is three basic principals on how Ethernet is a base band technology meaning only ONE signal can move down the line at one time (broadband sending 100`s television channels) works CS MA CD, Carrier Sense, first logic in the NIC uses carrier sense, it uses INT52, the NIC looks for digital pulses, if it sees pulses it wont send the data, if theirs no pulses and no one is sending zero data (volts) Carrier sense, no carrier zero volts, no one is sending things, it first makes sure nobody else is using the network, the second pillar of Ethernet wisdom is called multiple access, You can easily simply find the computer on the system , the third is the most important, (CSMA CD), it’s a simplistic design, Ethernet strategy the first work station, the first NIC that detects ugly pulses has a job to do it sends out a high frequency jamming signal, the NIC would try to send it again, Ethernets simplistic approach from this collision will wait on a RANDOM number, > DATA >T L > SA > DA, Zerox ThreeCOM started Ethernet, the original ThreeCOM internet message was like DATA . TYPE . SA. DA, what do you think type stands for…it’s a two bite field that represented which higher protocol, Ethernet has never worked alone it is a delivery boy, a delivery of data. Type (a number) sends the frame to another destination reads the data and looks at the type and understands to send the data to TCP data, it can be all sorts of different TYPE, logic when sending data take note of where its from look at the data and note it should be shot of to TCP program. (HUMAN RESOURCES) IEEE our friends did few things, started making standards which allowed different types of local area networks to communicate with each other. The IEEE came up with the standard 802, this started in 1980 in Feb, they said its going to be called 802.3 and token ring 802.5, 802.11 is wireless, one of the changes was the change of TYPE and made it length, why do you need a length field, well what if you were only sending 25 bytes of data that may or might be not be a important, how does the IEEE know what type of data..they borrowed the first two bites the IEEE solved the confusion of the none 802.3. Ex: Three companies Zerox, Digital Dix, the IEEE solved the problem, today they all worktogether,, not important for dcn 286 theirs a extra thing they put on the frame theirs a preamble, its 64 bits long, 101010101010101010101010 alternating ones and zeros. Think about this you’re a NIC that cost 25$ you can send info to each other, what would happen if you were in a bad mood, and u started sending smaller bits, how does this precision they get insync, very simple very basic that works. Primitive 10baseT 10 million bits persecond using base technology over twisted pair cables, think go Cat5 Cable, Lab wed Cat5 and make the cable straight threw and crossover. Beaware of this Cat5 comes int two types UTP and STP, unshielded twisted pair, and shielded twisted pair. Cat 5 6 7 learn to love next wed. READ up to chapter three.
Thursday, October 25, 2007
DCN 286 - Information on TCP....B/C You will be writing Essays On This ALL THE TIME!
What is ip and what does it do, what is ip’s job, Ip’s job is simple is to get our data through the internet as quickly as possible and that all its job is its job is not to make sure your sure data is safe Ip does not care about that and if it looses you’re data oh well, and not their for reliability all it cares about is trying its little best to send your data through the net as quickly as possible, ip runs not only in your computer, but in all the routers in the internet, client and server, all have ip, gets stuff done fast, you’re going to tell in week 14 about the total length (16 bits) it can theoretically take you’re data and put it into pieces 65,535 K the routers hate big packets, so if you accidently sent a big packet it would break it into fragements, if you approach a cisco router CISCOs default is 1500 bytes, you can change the router MTU bigger or smaller (maximum transmission unit) it uses the identification field so all the fragments have unique numbers, the router then takes a fragment offset value and see which position its on, middle first or last fragments so when all the fragments get to destination computer IP will reassemble the fragments in the original , TIME to live is how many hops a packet can go through the network, you can overwrite it depends on you’re OS, if the router can’t go to the destination it will get destroyed , last thing were going to tell him is the IP address, IP addressing 32 bits of total confusion, todays not ip but TCP, Tranfer Control Protocol, essays on TCP for the rest of you’re Seneca carrier, at first glance it looks very simple but very deceptive very complex, TCP provides reliability TCP/IP running on your computer, in the Internet cloud theirs huundred of thousands of routers, and all the routers are interconnected, the routers are also running IP when you pass the IP packet to the router and examines the destionation address. TCP is running in any server you connect to, what are you doing at home maybe you wanna download a picture of Pluto, www.nasa.org, scattered around the would there are DNS they have two database with NAMES AND ADDRESS, when you contact the DNS server send it a name and will return an IP address, TCP is going to provide realiability while you download the picture to do this it is going to use something called a port, a port when talking about TCP is a number a 16 bit number so 65, 535K, IEEE standards organization which came up with a list with internet application, DNS internet application, TELNET, FTP HTTP, anyone who, HTTP port 80, FTP port 21/22, TELNET 23 we are going to use FTP to download the picture of Pluto, the port numbers 0-1000, are reserved port numbers for legitimate internet application called Well known ports, over 1000 free to be used by any computer, Here you go you got the ip from dns and you know nasa, org TCP doesn’t know the address but attached to IP and its going to build a header information, default header size is 20 bytes and its going to start ot build port numbers destination port Destination port is 21, you can send lots of things to nasa.org you can try to telnet to nasa.org but you want to download something so you are using FTP, TCP will generate a random number will be between 1000, 65000, essentially its going to say this to the client machine to the server what your doing is TCP is opening a session between your client and nasa.org, there is now a session being established, to work correctly, when the seassion has been established the two machines can send data with eachother first open a session then send data then end the session, the session to be established though is a little bit complicated, a client machines sends segment to the server and will flip the two numbers around and send them back to you, here is the complicated part, TCP receives confirmation sends a third header switching it again, the two machines, when they are forming a session are negotiating with each other, before they need to send data, two or three things they must know, most important thing is how long does it takes TCP segment to go from here to the network and back again, the client initiates the second sending TCP record number 1 this machines looks at the system clock on the computer and record the time for the request of the session the little fragment goes flying through the network and sends it back and looks at the system clock again and now they know how long the round trip is going to know how long its going to take, lets just say it takes 1 second, but this machine does not know what it is and is the purpose of the third segment, is looks at the system clock, three headers, now both machines understands how long it takes to you’re computer and the server now it know it takes from it to you, other important things are being negotiated when tcp first came out, back when their was 50 computers and the internet tcp used sequence number and acknowledgement number for error checking, if there is no errors its going to send an acknowledgement number for error handling the sequence number in early days, use to start at sequence number 1 then 2,3(Frames) does not happened any , more, early days no hacking, sequence number was a hackers prayer, todays version of tcp, between the negotiation will pick another random number, you got 32 bits, what’s the maximum number 4 bil and change, so ip when its sending the initial request will pick a random number 0 -4bil and will stick it in the sequence number field (init sequence number) it will start counting at five when it gets it and when gets to the server another random number is chosen, to set this session we look at the little 1 bit of information when the client sends the first header its going to turn on the SYN (flag 1) when it sends the reply its going to turn on the ack bit (1) when it sends the third reply its going to turn of the syn(0) and send a reply, you notice tcp full duplex protocol, theirs multiple downloads, so lets see how it transits data, how weird is it? Where going to ftp to send data FTP, is a simple protocol, once the session is established it starts reading in data one byte at a time, when ftp thinks its got enough for one segment it will screen the data 1 byte of a time, and codes the bytes 1, 2, 3(bytes) if you left tcp alone it will leave it in your computer it would literally blow your memory of your computer, TCP says I think its enough data (100) bytes and knows the intial sequence number was 5 now its 105 and its going to do error checking on that 100 bytes, using Checksum an algorithm that makes sure theirs no bit errors and if there is no error bit its going to build the acknowledgement number 106, and turn on the acknowledgement number, then the server goes onto the next 100 bits and acknowledge number is going to go to 205, the data gets send tcp does error checking, the machine will build a twenty byte header and positive acknowledge number 300, if you peer closely of the little 1 bits their was a error, what would the server do? What tcp does is unique, TCP does nothing, TCP was written in such a way, TCP looks at the system clock and record the time, if theirs not a positive ack in 1 sec, tcp will automatically retransmit the packet, TCP will notice that the two packets with the same sequence number and discarded one of them, at the end of the session and all the data is sent it goes through a complex procedure to close down the session and we don’t have to worry about till next semester, Window the third thing the initial size, depending on the OS, link between tcp and the operating software, if your using Windows, windows will direct tcp will have an initial window size of 64K, Linux has much smaller window sizes 2k, the bigger the window size the easier you make the hackers job, when the session is nogiated it’s a initial window size, two tcp start sending data, lets say they start small window size of 2k, tcp will actually monitor the number of errors are their, if tcp ‘thinks’ the network is good today because of no errors either machines can increase the window size, and if even more it will increase, tcp increase and decrease the windows until errors start happening, window size is really how much memory is going to be reserved on each computer, the application determines how much or when to send data, it tells tcp heres enough data, and will turn on the push flag and could receive the blocks of bytes and save it in number, when a tcp receives a packet when its turn on it will saveit and push it into the application, its how much memory in each machine is going to be used, ultimately IP is going to see the packet size, Window size is the amount of memory to save multiple tcp pieces, Urgent Pointer and Urgent bit till next semester, for now all we have to know is TCP header and a big pile of data following the header, lets say for some reason the application wanted you to read a specific data that not right on the front it will turn on the urgent data and will read the password which was 100 bytes in, save till next semester. TCP and IP are mentioned numerous chapters into your computer, write essays in week 14 and put these two topics in test TWO to get some practice.
TCP
Source Port(1001) | Destionation Port (21)
Squence Number (00000101)
Acknowledgement Number
Data Offset| Reserved |URG(1)|ACK(0)|PSH|RST|SYN|FIN | window
Checksum | Urgent Pointer(1)
Options + Padding
VERSION IHL TYPE OF SERVICE TOTAL LENGTH
What is a collusion domain, HUB with ports, you got a HUB the simultaneously the collusion will interfere with all the toerh computer, and the chaos is extened to every machines (CD) and the more machines the more chance of collusions happening, Switches are employeed to reduce the effect of collusions of your network, switches break it up to separate collision domains, SWITCH 5 ports and traditional little network hubs to each port and then attached to networks, no the hubs machines are in a collusion domain but if a collusion hapends here and will have no effect on the other domain, Switches reduce the collisions, everything we put on here is its own little collision domain, today happy networking use a process called micro segmentation, what does that mean, networking guys today have virtually eliminated collision through high quality switches and having one machine on each port on a switch their can be no collisions , micro segmentation imply switches and are all into switches Cut THREW switch does what, when a cut threw switch reads the frame it immediately checks the destination address, will immediately set up a channel source and destination port, if you got a nice safe environment, Seneca college, you use cut threw switches, how ever if the environment is a little shakey, in that environment you use STORE and FORWARD switch that has huge amounts of memory and will read the face and its entirety and store the frame where it does error checking, and if the frame takes it and there is no errors it will send it to its destination,
Does not matter which but it can get to one work station through its port, now we want to talk about subnets, switches are datalink layer devices routers are network layer devices, we are one originations and we sell stuff on the net the switch to the computer and the server, we need to sell stuff on the net so we need to buy a router, we need a domain name, our domain name has a associated ip address 200.1.1.0 (ClassC) we hire a networking switch 200.1.1.254, we got a problem cause were in Toronto and going to Vancouver, to they subnet to create it so they can expand and have their own portion, to create subnets its an easy thing don’t matter what class, you create them the same way you play around with binary numbers,windows create subnets a dumb way, and use a short cut and don’t know when there is a lot of subnets, the correct way is to play with binary numbers, which numbers
Network Address 200.1.1.[0] in control of the 4th octet, to create subnets take the host bits and write them out in binary one rule to be applied to subnets by convention we never use all zeros or ones, so we have two left to indicated subnets 00 01 10 11 so 01 becomes Toronto while 10 becomes Vancouver. [00]000000 to indicate if its Toronto or vancouver
SN
00|000000
01|000001 [65]1*
01|111110 [126]2*7th -2
11|111111
We turn on the host bits and the subnet
255.255.255.192 which is the value of [11]000000
1* the address can be added for example 010000010 and moving the 1 up to the numbers
The last number and the range there for becomes 65 – 126
Vancouver
SN
10|000001[129]
10|111110[190]
NOW WE NEED THREE SUBNET IDS with 001 = T. 010 = V 100=LA
001|00001 [33]
001|11110 [62]
TCP
Source Port(1001) | Destionation Port (21)
Squence Number (00000101)
Acknowledgement Number
Data Offset| Reserved |URG(1)|ACK(0)|PSH|RST|SYN|FIN | window
Checksum | Urgent Pointer(1)
Options + Padding
VERSION IHL TYPE OF SERVICE TOTAL LENGTH
What is a collusion domain, HUB with ports, you got a HUB the simultaneously the collusion will interfere with all the toerh computer, and the chaos is extened to every machines (CD) and the more machines the more chance of collusions happening, Switches are employeed to reduce the effect of collusions of your network, switches break it up to separate collision domains, SWITCH 5 ports and traditional little network hubs to each port and then attached to networks, no the hubs machines are in a collusion domain but if a collusion hapends here and will have no effect on the other domain, Switches reduce the collisions, everything we put on here is its own little collision domain, today happy networking use a process called micro segmentation, what does that mean, networking guys today have virtually eliminated collision through high quality switches and having one machine on each port on a switch their can be no collisions , micro segmentation imply switches and are all into switches Cut THREW switch does what, when a cut threw switch reads the frame it immediately checks the destination address, will immediately set up a channel source and destination port, if you got a nice safe environment, Seneca college, you use cut threw switches, how ever if the environment is a little shakey, in that environment you use STORE and FORWARD switch that has huge amounts of memory and will read the face and its entirety and store the frame where it does error checking, and if the frame takes it and there is no errors it will send it to its destination,
Does not matter which but it can get to one work station through its port, now we want to talk about subnets, switches are datalink layer devices routers are network layer devices, we are one originations and we sell stuff on the net the switch to the computer and the server, we need to sell stuff on the net so we need to buy a router, we need a domain name, our domain name has a associated ip address 200.1.1.0 (ClassC) we hire a networking switch 200.1.1.254, we got a problem cause were in Toronto and going to Vancouver, to they subnet to create it so they can expand and have their own portion, to create subnets its an easy thing don’t matter what class, you create them the same way you play around with binary numbers,windows create subnets a dumb way, and use a short cut and don’t know when there is a lot of subnets, the correct way is to play with binary numbers, which numbers
Network Address 200.1.1.[0] in control of the 4th octet, to create subnets take the host bits and write them out in binary one rule to be applied to subnets by convention we never use all zeros or ones, so we have two left to indicated subnets 00 01 10 11 so 01 becomes Toronto while 10 becomes Vancouver. [00]000000 to indicate if its Toronto or vancouver
SN
00|000000
01|000001 [65]1*
01|111110 [126]2*7th -2
11|111111
We turn on the host bits and the subnet
255.255.255.192 which is the value of [11]000000
1* the address can be added for example 010000010 and moving the 1 up to the numbers
The last number and the range there for becomes 65 – 126
Vancouver
SN
10|000001[129]
10|111110[190]
NOW WE NEED THREE SUBNET IDS with 001 = T. 010 = V 100=LA
001|00001 [33]
001|11110 [62]
DCN286 - Information on IP...GOOD INFO
Browsers are not internet applications,
HTTP, TELNET, FTP, DNS --- APS
All internet applications were written so they can link with the transport protocol, all these applications work with TranferControlProtocol and UserDatagramProtocol , you can do a lot of things with this write C++, Two guys wrote YouTube.com they used an idea that anyone in the room can do, they wrote an internet application that was very successful, the science behind it is common, DCN386 writing an internet application program…they use a ApplicationProgramInterface…this tool will hook this application right into TCP or UDP, TCP what does it do? Tcp does very little actually it provides reliability into your internet transmission, here is beautiful drawing of the internet CLOUD with a bunch of router drawings along with you on the side at home on the net…and on the left the ISP….The first thing you have to know about the internet and how ti works when you send packets of data in to the internet you sending it into a hostile environment…it is dangerous inside the cloud…ROUTERS which we think is a our best friend…it turns Routers hate packets they detest them they only get rid of them, but the easiest way is to move it to the next router…because its such a hostile environment TCP running in your computer and the SERVER in which your contacting TCP tries its best when Routers destroy you packet TCP will contact the server and ask it to retransmit the packets…TCP will keep trying and keep trying a router does not know anything about retransmitting stuff…so TCP just runs in servers or clients it doesn’t run in the internet its reliability in each end. UDP on the other hand is a stripped down version or TCP…error handling is not in that UDP is unreliable while TCP is reliable…Why would they use the unreliable and a reliable, UDP faster then TCP, depending on what you want to do…the rule says if you sending tiny packets into the network…use UDP because the chances of an error is small so send your tiny packet through the net as fast as you can…ok lets look at a lot of a big picture…going to nasa.com to get a picture of Pluto….nasa.org will create DomainMain(its Name(lol))Service…global system where you type in nasa.org and it will go find NASA’s ip address…so your typing it in and DNS will treat it as data and will look at the data to the closest DNS server….its a tiny amount of data…What do you think DNS uses…obviously UDP…DNS always uses UDP…okay you got the ip addresss… TCP will pass to a program called IP and will send info to nasa.org that has a big file..pluto.jpeg now that’s big files and its going to be downloaded using FTP its going to download a lot of information…big packets are involved…what do you think FTP is going to do? FTP is running on the server nasa.org…it read the file and sends it into the internet FTP always interact with TCP because it doesn’t want any mistakes these two programs only run on the CLIENT or SERVER machine they have to no part into the Internet…it just knows reliability. How does TCP get itself through the network it will read the |data|F(header file)|TCP puts a lot of control information on it 20 byte header to be exact| this then passes down to IP everything uses IP…this get passed down |data|TCP|IP(20bytes of control information|), both have 20 bytes of information its 40 bytes of information must learn all 40 bytes for test two…TCP can deliver anywhere in the world…IP doesn’t know how to get into the cloud its great when its inside the cloud but it has no clue on how to get in their, what does IP do it sends it down to 802.3, NIC you put in it’s a Ethernet NIC…Nic will send it and you ISP is inside the cloud. Ethernet is going to deliver that packet from your computer to your ISP…whats in the Ethernet |data|TCP|IP|Ethernet … when the ISP gets the packet it gets read of the header and strips it off and passes directly to the Router inside the cloud Ethernet delivers the data. Now the IP packets inside the cloud it can be delivered to the destination Address..when it gets close to the router where the information is …its forwards through the routers and before it goes to the ISP and is re-stripped from the Ethernet frame. The internet is a popular their. IP we have to figure out how IP works….first thing we have to know about IP IP like routers does not care about your data…it doesn’t careless if you packet makes it or if it doesn’t. IP makes the best effort through the network as quicklWy as possible…HOSTILE. Where going o look at 20 bytes of infor
0 4 8 16 19 31
| Version | IHL | Type of service | Total Length |
| Identification | Flags | Fragment offset|
| Time to live | Protocol Header | CheckSum |
| Source Address |
| Destination Address |
| Options + padding |
The very first bits look at Version it will look at 0100 the binary level which equals 4 it will all have four bits of information…sooner or later your ISP is going to be version 0110 version 6, a couple of years ago he would say version six will be introduced right away…the IP address is 32 bytes long…the classes made it less effective and proved to be not a great idea and ran out , for a start the IP header has 128 bits, when IP version six comes out, he can go around the world in one square feet increments, in version six each square foot will have million in the square foot…the reason were not panicking anymore came up with something Class List Ip ADDRESSING because of that the full amount of IP address can now be access and now it is being revised…somebody’s brain wave made v4 not needing v6 but Linux. Their will be a compatibility with each other v4 and v6. The next four bits are equally boring Intial Header Length….(20 octets (20 bytes)) if we were creating are own internet application and wanted to test it out we would use something called options…Options is a little bundle of tools that will let us test are application…never do unless you writing your own internet application…..Ever heard of a little application Python…he took look of Python and thought of us…Python already has (application program tool kit to write applications) Python just installed in the Cisco lab…make your own client and make your server…...Type of service.. IP is probably the same age of us it was written quite a while ago..its been around for a couple of decades for 20 years all routers in the internet ignored the 8 bytes…all of a sudden few years ago its become the hottest 8 bits Quality of Service in Networking, we got to worry about the 8 bits…something happened recently another internet ..downloading Movies Voice and Media…Type of Service came important some packets are going to have priority of other packets. Router has serial ports and Ethernet ports…packets come into the router… routers temporarily store IP packets were talking nano seconds it just buffers them in old days they would look at ToS but now routers are starting to examine this Type of Service, this particular packet is in the middle of the buffer but has quality service and will process it first…just understand that some routers can treat some packets as priority packets…Total Length…that means that Total length for Data|TCP|IP the data and the control information ..Ponder how big can an IP packet can be 65536 (64 k) that is a big packet…64 k packet if you were foolish enough to write a program to send out 64 k you would be sorry…. Routers hate big packets they despise… they wont destroy them they will take your big packet and chop it into little fragments of the original packet… all routers do this… hard to comprehend.. when we go to cisco lab on wed and we look at the router and start configuring the router they have a default maximum tramission unit size… if it receives a big packet it will be MTU’s its 1500 bytes… because that’s what Ethernet uses….Lets drop a nice big packet | 4500(chop into three) |(20 byte header)
| (1500)[ID010] | (1500) [ID100]| (1500) [ID001] | - contained in the header info is the destination it chops up and put the header infront of each fragment.
How does the server know that they all belong together because the server gets all the fragments has to put them back together HOW DOES IT KNOW the three fields of identification flags and fragments offsetthis is what indicates the fragments belong together…IP before it sent the packet it picked a random number and put it in the Identification field you’ll notice it will be 16 bits its going going to be 0 – 65535 bytes… when the router breaks the big thing into pieces each fragment will have the same identification number and knows they are fragments and have to be reassembled back into the frame ,….wait theirs a problem went different ways…How does IP manage to reassemble these three fragments…they would look at the fragment offset...
|5000 |3000 |1500 | |byte 0 its going to read them and store them temporarily in memory…it takes the data read the offset and puts it in memory but by the OFFSET of 3000… it takes the 5000 and puts it the 5000 OFFSET…sequence failed, what if the sequence 1 get into the a grumpy router fragments can get fragmented….The Flags are three bits long…only two of the bits are used the first bit is a strange bit.. if the first bit is turned on that means that ip examining this it means DO NOT FRAGMENT its called a do not fragment bit…it was originally designed to send out and see if you get anything back… if a router got that packet with the do not fragment it will destroy it… the second bit means Last Fragment Indicator, this is used so after it knows it has all the fragments because of this bit. It knows it has the last one in the series…8bits Time to LIVE….what does TCP/IP care about to time to live… what if we purposely put a destination address that does not exist, theoretically our packet can go through the world…Time to live when your IP sent the packet out it put a value in TTL it will be some value from windows to LINUX from 1 -255 and the default value is 32bit, it will subtract the value and it will eventually go to zero with the zero packet will destroy it….sending a packet out of Seneca it probably …www.bagpiprepaire.com do a trace route and you’ll see it…. Protocol not a very important field it simply means TCP create the this packet or did UDP create this packet, and will use a different code for each so when the packet delivers the packet it look at it and will be like deliver it to TCP or IP. code maintained by the IEEE, IP can run with a lot of other protocols…Raw sockets, you just use IP on its own… third code….IP does not care about your data….Header Checksum.. Does IP do error checking…yes it does error checking only on the Twenty bits and only looks at the the IP address….WE have to totally understand the IP address because in test 2 we are going do questions called Subnet ting questions and to do Subnet ting we have to understand the IP address really really well. Early start on subnetting. Guys that created the IP standard, origanilly created class addressing… what can you tell to determine the class of the address you simple look at the first byte. IP addresses are four bytes long.. class a determined looking at the first byte they all have a zero of the first bits of the first bytes leaving seven bits , the other three are Host address and class B are 127 (127 is special called LOOPBACK. Used for testing) B 10111111 128-191 C 11011111 192-223, How many class a’s 126, 65535kb, class C 254 hosts. Working in groups of two ooo, we want first stations to the other station through two routers what have to do to achieve this is do something configuration file, basic configurations.. apply IP address to the PORTS we want to apply ip address to those ports.. they will learn how to send stuff through a network… all we do is provide the initial information and press on the routers will figure out what to do, Lets say have an ip network and were going to use a class a address 10.0.0.0 that number means the entire network but to a router they don’t see one big network but THREE subnets, how do we break it into three subnets we are going to use a Subnet Masks, this allows you to create these subnets in the first place… you see if we just had the network number and we plug them in we wouldn’t know which router can get the total packet.. so we use the mask 255.255.0.0 if for some reason we didn’t wanna create subnets we could give an ip address 10.0.0.1 we have to put a 10.0.0.1, 2, 3, 4, 5, 6 in each port, things get broken into subnets…we wouldn’t do it without the subnet and we would still have to have a mask and have a default mask 255.0.0.0 a one byte means this bits parts of the network ID if we wanted to know subnetting we take the second byte and will be used to indicate the bits are subnet bits hence if we wanted to do that we make 255.255.0.0 11111111.11111111.0.0.0, we will change the purpose of the bits and when the router recivies the packets will receive that packets one being the counter and one the subnet….Look up subnetting…
HTTP, TELNET, FTP, DNS --- APS
All internet applications were written so they can link with the transport protocol, all these applications work with TranferControlProtocol and UserDatagramProtocol , you can do a lot of things with this write C++, Two guys wrote YouTube.com they used an idea that anyone in the room can do, they wrote an internet application that was very successful, the science behind it is common, DCN386 writing an internet application program…they use a ApplicationProgramInterface…this tool will hook this application right into TCP or UDP, TCP what does it do? Tcp does very little actually it provides reliability into your internet transmission, here is beautiful drawing of the internet CLOUD with a bunch of router drawings along with you on the side at home on the net…and on the left the ISP….The first thing you have to know about the internet and how ti works when you send packets of data in to the internet you sending it into a hostile environment…it is dangerous inside the cloud…ROUTERS which we think is a our best friend…it turns Routers hate packets they detest them they only get rid of them, but the easiest way is to move it to the next router…because its such a hostile environment TCP running in your computer and the SERVER in which your contacting TCP tries its best when Routers destroy you packet TCP will contact the server and ask it to retransmit the packets…TCP will keep trying and keep trying a router does not know anything about retransmitting stuff…so TCP just runs in servers or clients it doesn’t run in the internet its reliability in each end. UDP on the other hand is a stripped down version or TCP…error handling is not in that UDP is unreliable while TCP is reliable…Why would they use the unreliable and a reliable, UDP faster then TCP, depending on what you want to do…the rule says if you sending tiny packets into the network…use UDP because the chances of an error is small so send your tiny packet through the net as fast as you can…ok lets look at a lot of a big picture…going to nasa.com to get a picture of Pluto….nasa.org will create DomainMain(its Name(lol))Service…global system where you type in nasa.org and it will go find NASA’s ip address…so your typing it in and DNS will treat it as data and will look at the data to the closest DNS server….its a tiny amount of data…What do you think DNS uses…obviously UDP…DNS always uses UDP…okay you got the ip addresss… TCP will pass to a program called IP and will send info to nasa.org that has a big file..pluto.jpeg now that’s big files and its going to be downloaded using FTP its going to download a lot of information…big packets are involved…what do you think FTP is going to do? FTP is running on the server nasa.org…it read the file and sends it into the internet FTP always interact with TCP because it doesn’t want any mistakes these two programs only run on the CLIENT or SERVER machine they have to no part into the Internet…it just knows reliability. How does TCP get itself through the network it will read the |data|F(header file)|TCP puts a lot of control information on it 20 byte header to be exact| this then passes down to IP everything uses IP…this get passed down |data|TCP|IP(20bytes of control information|), both have 20 bytes of information its 40 bytes of information must learn all 40 bytes for test two…TCP can deliver anywhere in the world…IP doesn’t know how to get into the cloud its great when its inside the cloud but it has no clue on how to get in their, what does IP do it sends it down to 802.3, NIC you put in it’s a Ethernet NIC…Nic will send it and you ISP is inside the cloud. Ethernet is going to deliver that packet from your computer to your ISP…whats in the Ethernet |data|TCP|IP|Ethernet … when the ISP gets the packet it gets read of the header and strips it off and passes directly to the Router inside the cloud Ethernet delivers the data. Now the IP packets inside the cloud it can be delivered to the destination Address..when it gets close to the router where the information is …its forwards through the routers and before it goes to the ISP and is re-stripped from the Ethernet frame. The internet is a popular their. IP we have to figure out how IP works….first thing we have to know about IP IP like routers does not care about your data…it doesn’t careless if you packet makes it or if it doesn’t. IP makes the best effort through the network as quicklWy as possible…HOSTILE. Where going o look at 20 bytes of infor
0 4 8 16 19 31
| Version | IHL | Type of service | Total Length |
| Identification | Flags | Fragment offset|
| Time to live | Protocol Header | CheckSum |
| Source Address |
| Destination Address |
| Options + padding |
The very first bits look at Version it will look at 0100 the binary level which equals 4 it will all have four bits of information…sooner or later your ISP is going to be version 0110 version 6, a couple of years ago he would say version six will be introduced right away…the IP address is 32 bytes long…the classes made it less effective and proved to be not a great idea and ran out , for a start the IP header has 128 bits, when IP version six comes out, he can go around the world in one square feet increments, in version six each square foot will have million in the square foot…the reason were not panicking anymore came up with something Class List Ip ADDRESSING because of that the full amount of IP address can now be access and now it is being revised…somebody’s brain wave made v4 not needing v6 but Linux. Their will be a compatibility with each other v4 and v6. The next four bits are equally boring Intial Header Length….(20 octets (20 bytes)) if we were creating are own internet application and wanted to test it out we would use something called options…Options is a little bundle of tools that will let us test are application…never do unless you writing your own internet application…..Ever heard of a little application Python…he took look of Python and thought of us…Python already has (application program tool kit to write applications) Python just installed in the Cisco lab…make your own client and make your server…...Type of service.. IP is probably the same age of us it was written quite a while ago..its been around for a couple of decades for 20 years all routers in the internet ignored the 8 bytes…all of a sudden few years ago its become the hottest 8 bits Quality of Service in Networking, we got to worry about the 8 bits…something happened recently another internet ..downloading Movies Voice and Media…Type of Service came important some packets are going to have priority of other packets. Router has serial ports and Ethernet ports…packets come into the router… routers temporarily store IP packets were talking nano seconds it just buffers them in old days they would look at ToS but now routers are starting to examine this Type of Service, this particular packet is in the middle of the buffer but has quality service and will process it first…just understand that some routers can treat some packets as priority packets…Total Length…that means that Total length for Data|TCP|IP the data and the control information ..Ponder how big can an IP packet can be 65536 (64 k) that is a big packet…64 k packet if you were foolish enough to write a program to send out 64 k you would be sorry…. Routers hate big packets they despise… they wont destroy them they will take your big packet and chop it into little fragments of the original packet… all routers do this… hard to comprehend.. when we go to cisco lab on wed and we look at the router and start configuring the router they have a default maximum tramission unit size… if it receives a big packet it will be MTU’s its 1500 bytes… because that’s what Ethernet uses….Lets drop a nice big packet | 4500(chop into three) |(20 byte header)
| (1500)[ID010] | (1500) [ID100]| (1500) [ID001] | - contained in the header info is the destination it chops up and put the header infront of each fragment.
How does the server know that they all belong together because the server gets all the fragments has to put them back together HOW DOES IT KNOW the three fields of identification flags and fragments offsetthis is what indicates the fragments belong together…IP before it sent the packet it picked a random number and put it in the Identification field you’ll notice it will be 16 bits its going going to be 0 – 65535 bytes… when the router breaks the big thing into pieces each fragment will have the same identification number and knows they are fragments and have to be reassembled back into the frame ,….wait theirs a problem went different ways…How does IP manage to reassemble these three fragments…they would look at the fragment offset...
|5000 |3000 |1500 | |byte 0 its going to read them and store them temporarily in memory…it takes the data read the offset and puts it in memory but by the OFFSET of 3000… it takes the 5000 and puts it the 5000 OFFSET…sequence failed, what if the sequence 1 get into the a grumpy router fragments can get fragmented….The Flags are three bits long…only two of the bits are used the first bit is a strange bit.. if the first bit is turned on that means that ip examining this it means DO NOT FRAGMENT its called a do not fragment bit…it was originally designed to send out and see if you get anything back… if a router got that packet with the do not fragment it will destroy it… the second bit means Last Fragment Indicator, this is used so after it knows it has all the fragments because of this bit. It knows it has the last one in the series…8bits Time to LIVE….what does TCP/IP care about to time to live… what if we purposely put a destination address that does not exist, theoretically our packet can go through the world…Time to live when your IP sent the packet out it put a value in TTL it will be some value from windows to LINUX from 1 -255 and the default value is 32bit, it will subtract the value and it will eventually go to zero with the zero packet will destroy it….sending a packet out of Seneca it probably …www.bagpiprepaire.com do a trace route and you’ll see it…. Protocol not a very important field it simply means TCP create the this packet or did UDP create this packet, and will use a different code for each so when the packet delivers the packet it look at it and will be like deliver it to TCP or IP. code maintained by the IEEE, IP can run with a lot of other protocols…Raw sockets, you just use IP on its own… third code….IP does not care about your data….Header Checksum.. Does IP do error checking…yes it does error checking only on the Twenty bits and only looks at the the IP address….WE have to totally understand the IP address because in test 2 we are going do questions called Subnet ting questions and to do Subnet ting we have to understand the IP address really really well. Early start on subnetting. Guys that created the IP standard, origanilly created class addressing… what can you tell to determine the class of the address you simple look at the first byte. IP addresses are four bytes long.. class a determined looking at the first byte they all have a zero of the first bits of the first bytes leaving seven bits , the other three are Host address and class B are 127 (127 is special called LOOPBACK. Used for testing) B 10111111 128-191 C 11011111 192-223, How many class a’s 126, 65535kb, class C 254 hosts. Working in groups of two ooo, we want first stations to the other station through two routers what have to do to achieve this is do something configuration file, basic configurations.. apply IP address to the PORTS we want to apply ip address to those ports.. they will learn how to send stuff through a network… all we do is provide the initial information and press on the routers will figure out what to do, Lets say have an ip network and were going to use a class a address 10.0.0.0 that number means the entire network but to a router they don’t see one big network but THREE subnets, how do we break it into three subnets we are going to use a Subnet Masks, this allows you to create these subnets in the first place… you see if we just had the network number and we plug them in we wouldn’t know which router can get the total packet.. so we use the mask 255.255.0.0 if for some reason we didn’t wanna create subnets we could give an ip address 10.0.0.1 we have to put a 10.0.0.1, 2, 3, 4, 5, 6 in each port, things get broken into subnets…we wouldn’t do it without the subnet and we would still have to have a mask and have a default mask 255.0.0.0 a one byte means this bits parts of the network ID if we wanted to know subnetting we take the second byte and will be used to indicate the bits are subnet bits hence if we wanted to do that we make 255.255.0.0 11111111.11111111.0.0.0, we will change the purpose of the bits and when the router recivies the packets will receive that packets one being the counter and one the subnet….Look up subnetting…
Subscribe to:
Posts (Atom)