Friday, November 2, 2007

DCN386W9 - SDL HDLC FRAMERELAY RIPv1

Question 1 Test 2

Test 2 in week 13

Scummed the test AAAAAAAAAHHHHHHHHHHHHHHHRRRRRRRRHHHHHHHHHH!


In the course outline
2test each worth 25%,
1Q 10%
FE 40%

Test 1 20%


60/3




Question 1

It’s a protocol

SDLC- Synchronous Data link Control Protocol

For the 100$ coming out of your bank accounts, you use SDLC

IBM’s protocol very expensive

OSI changed SDCL to –HDLC-

HDLC is ciscos default datalink protocol, if you understand SDLC you will understand HDLC

This is a data link layer protocol
IETF renamed it to PPP, is a derivative of HDLC which is the DERATIVE of SDLC

SDLC insights to half a dozen other modern protocols, SDLC and HDLC and PPP and FrameRelay are lay DATALINK layer protocols,

How do you catergorize it, datalink protocols are PPP protocols

What does this mean exactly,


You and to the bank is PPP no Network Layer Involved, by week 14 why is is not a good idea to put a bank machine on the Internet.

You don’t have to worry about Ordered Delivery theirs only one route

It is stressed it is a Data link protocol,

When you want to learn about protocol you look at the control information that is infront TCP/IP 20 bytes of information in front of every protocol.

He will send us a website with the information to the diagram from this lecture

SDLC + HDLC + PPP and all the things have a
Variable data size – the protocol does not specify the maximum data unit

Ethernet is also a data link protocol it has a maximum variable data size.

If you worked at Scotia bank you would pick the maximum data size, it has versatility you can do what you want with the variable size.

U would say if setting up a bank application down at the bottom of the script would be
MAXDATA : 100

The address 8 bits, 2 bytes long

Lets go through this you ask for 100 bucks from a bank machine it will put the information into the data field, the bank machines surrounds your data with this extra information, a 1 byte
FLAG – 8 bit of information - 01111110
Frame Relay same information it starts with a flag

Why do we need flags, know when we know the message is over.

The reason you need a flag cause SDLC does not have a fixed data size,

Ex: what happends a receiving machines gets a start flag it reads data, onlt when it sees an end flag it knows it has received a complete message.

Start flag and end flag are like book marks between the data.

SDLC will guard against theirs another 6 bits that our not ours, when they all want to send a message the protocols will examines the data, looking for an occurrence of six consecutive ones

While building a frame the data is examined if the software (HDLC) if it ever finds 5 consecutive 1 bits it will stick an artificial zero at the end it will say AH I have a astart frame the computer will then remove the artificial bit and restore the bit, this is called BIT stuffing after 5 consecutive 1’s.

The 1’s are used cause we don’t know the size of the data field

The address is manually applied, one thing to consider, a 1 byte control field,

TCP/IP had 20 bytes of control information, very very inefficient, SDLC 8 bits of control information these 8 bits DETERMINE how SDLC operates, To control 8 bits wonderfully efficiently.


All the protocols we we are going to look at have
THREE FRAME formats its determined by the CONTROL FIELD


This is the first format, theirs a 0 in the control field that means an

information frame is being transmitted, however if if its 1 then a 0 then we are dealing with a

supervisory frame – positive ack, neg ack, and flow control feature

We are doing this with 8 bits instead of 20 bytes like TCP

So we got sending data, error handeling done by the supervisory frame,

A THIRD FRAME

Unnumbered – Used for Network Management Functions

its specially special, Network Management software, theirs number frames with 1 command u can take down and put up

SDLC uses what is called a SLIDING WINDOW (this is the window size), they call this a

Sequence Number
000 – Received Sequence Number
000- Send Sequence Numbers

SDLC can send 7 frames before an acknowledgement is demanded
Totally Reliable, it’s a VERY VERY.

Why 7 frames?

001
010
011
It will senevn frames cause the SN only has 7 frames cause it only has 3 bits,
The seven frames will go out into the internet and the deatination computer will get the 7 frames and then it will do
Error Checking
Under 7 frames CRC on each and every frames
If they are all correct the receiving computer will build a supervisory frame in the reciving computer it will put receive data
111

Function Code
00 means no errors
Supervisors address control and flags

The sending machine sends the other information, once it gets to 7 gets back to zero then starts again

What if their was an error in frame 6, it does CRC error checking.

The receiving computer sends a supervisory frame of a function code of 01 indicating their has been an error, 110.


Supervisory from 11 means stop sending data

Supervisory three things, pos ack, neg ack , and flow control

Whether its an information , supervisory and unnumbered frame, why is it unnumbered, cause they have received sequence number
Information have reveinumber send number

Unnumbered have 5 bits of function code, bring a network down up , reset the machine, 5 bits has 32 different functions, Network Manage Person can do function

Understand unnumbered frams perform network management functions.

Supervisory frames never send data, address control and end flag,

You can send data unnumbered frames, buy your RRSP the NMPerson, here’s a little

The final bit

Pull Final – 1 bit has 2 jobs, depends on which direction the message is moving, with SDLC central computer are called primary units and others primary units, primary units are computers that control the ordered delivery of data

Their can be no collisions in SDLC, think about it,

There is two people and they are side by side and they both ask for $100 and press enter at exactly the same time what will happen

POLLING is used for flow control, hence the name POLL, the bank machine handling SDLC each circuit SDLC will send a POLLING message is a supervisory frame with no data the bit in the middle will be turned on,

It will send it to the new frame ask if it has data pull back and see if it needs pulls the data then it goes to the next machine,

POLLING to each machine, pul indicator primary to secondary

Different if its coming from the secondary unit to the primary unit its called a final indicator bit.

Final frame it know it has received 1 2 3 and there is no more frames, 1 bit with 2 jobs.




Next 5 weeks:

WAN protocol, and Internet protocol, Test 2 WAN protocol questions and Internet specific questions

When you typed router rip
Router RIP- To build something called a Routing Information Table

Two flavours of routing information protocols RIP is the oldest.
RIP V2

Configuring router 1, R1 knows two things, R1 knows about subnet 1 and subnet 2,
R2 knows about S2 S3, R3 S3 S4.

RIP starts up in your router, does something intresting, will send out ever interface it knows and every serial it knows


RIP REQUEST – frame of data – 2 pieces of information, it knows about S1 and S2 and will send it to its neighbour, when R2 gets the rip request will send a rip request, it will say it knows about Subnet3 and Router 3
R1
IP address | Hop Count | Interface
S3 1 S0
*Add*
S4 2 S0

Now R1 knows that R1 is associated with R2 subnet 3

R2
IP address | Hop Count | Interface
S1 1 S1
S4 1 S0



Every 20 seconds when started, biggest problem, every 20 seconds on average it sends a RIP request to its Neighbours.

Another 20 seconds pass so R2 will send everything it knows ^^Addings to RIP: router 1 will *Add*

Second Major problem

IETF webpage, - Specification for maximum Hop count of 15

If you connect 2 routers together it will see that it will be 1hop count, rip will always send it out the fasts, the minimum HOP count.

Next week A FAR BETTAH routing PROTOCOL HA HA HA AHHHHH!

OPEN SORTING PATH is by far a superior protocol, far more complex protocol.

When you type in router RIP its rip version rip, the third crucial failure it cannot understand subnets. RIPv2 does address subnets

No comments: