Scribd
Upload
41 views

CSE 461: Bits and Bandwidth: Next Topic

This document discusses how data is transmitted across physical links. It covers different types of media like twisted pair, fiber and wireless; encoding schemes to represent bits as signals; clock recovery methods; and framing to delineate messages. The key parameters of a link are its bandwidth, which determines transmission time, and propagation delay, which along with transmission time determines overall latency.

Uploaded by

Shahban ktk
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
41 views

CSE 461: Bits and Bandwidth: Next Topic

This document discusses how data is transmitted across physical links. It covers different types of media like twisted pair, fiber and wireless; encoding schemes to represent bits as signals; clock recovery methods; and framing to delineate messages. The key parameters of a link are its bandwidth, which determines transmission time, and propagation delay, which along with transmission time determines overall latency.

Uploaded by

Shahban ktk
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

CSE 461: Bits and Bandwidth

Next Topic

 Focus: How do we send a message


across a wire?
Application
 The physical / link layers: Presentation
1. Different kinds of media Session
2. Encoding bits, messages Transport
3. Model of a link Network
Data Link
Physical
1. Different kinds of media
 Wire
 Twisted pair, e.g., CAT5 UTP, 10  100Mbps, 100m
 Coaxial cable, e.g, thin-net, 10  100Mbps, 200m
 Fiber
 Multi-mode, 100Mbps, 2km
 Single mode, 100  2400 Mbps, 40km
 Wireless
 Infra-red, e.g., IRDA, ~1Mbps
 RF, e.g., 802.11 wireless LANs, Bluetooth (2.4GHz)
 Microwave, satellite, cell phones, …

Fiber
 Long, thin, pure strand of glass
 light propagated with total internal reflection
 enormous bandwidth available (terabits)

Light source Light detector


(LED, laser) (photodiode)

 Multi-mode allows many different paths, limited by dispersion


Aside: bandwidth of a channel

 EE: bandwidth (B, in Hz) is the width of the pass-band in


the frequency domain
 CS: bandwidth (bps) is the information carrying capacity
(C) of the channel

 Shannon showed how they are related by noise


 noise limits how many signal levels we can safely
distinguish
 geekspeak: “cannot distinguish the signal from the
noise”

2. Encoding Bits with Signals

 Generate analog waveform (e.g., voltage) from digital


data at transmitter and sample to recover at receiver

1
0

 We send/recover symbols that are mapped to bits


 Signal transition rate = baud rate, versus bit rate
NRZ and NRZI

 Simplest encoding, NRZ (Non-return to zero)


 Use high/low voltages, e.g., high = 1, low = 0
 Variation, NRZI (NRZ, invert on 1)
 Use transition for 1s, no transition for 0s

Bits 0 0 1 0 1 1 1 1 0 1 0 0 0 0 1 0

NRZ

Clock Recovery

 Problem: How do we distinguish consecutive 0s or 1s?


 If we sample at the wrong time we get garbage …
 If sender and receiver have exact clocks no problem
 But in practice they drift slowly
 This is the problem of clock recovery

 Possible solutions:
 Send separate clock signal  expensive
 Keep messages short  limits data rate
 Embed clock signal in data signal  other codes
Manchester Coding
 Make transition in the middle of every bit period
 Low-to-high is 0; high-to-low is 1
 Signal rate is twice the bit rate
 Used on 10 Mbps Ethernet

 Advantage: self-clocking
 clock is embedded in signal, and we re-sync with a
phase-locked loop every bit
 Disadvantage: 50% efficiency

Coding Examples

Bits 0 0 1 0 1 1 1 1 0 1 0 0 0 0 1 0

NRZ

Clock

Manchester

NRZI
4B/5B Codes

 We want transitions *and* efficiency …


 Solution: map data bits (which may lack transitions) into
code bits (which are guaranteed to have them)

 4B/5B code:
 0000  11110, 0001  01001, … 1111  11101
 Never more than three consecutive 0s back-to-back
 80% efficiency

 This code is used by LANs such as FDDI

3. Framing

 Need to send message, not just bits


 Requires that we synchronize on the start of
message reception at the far end of the link
 Complete Link layer messages are called frames

 Common approach: Sentinels


 Look for special control code that marks start of
frame
 And escape or “stuff” this code within the data
region
Example: Point-to-Point Protocol
(PPP)
 IETF standard, used for dialup and leased lines

Flag Payload Flag


(header) (trailer)
0x7E (variable) 0x7E

 Flag is special and indicates start/end of frame


 Occurrences of flag inside payload must be “stuffed”
 Like an “escape” character:
• \.\.\”\\ --> ..”\
 Replace 0x7E with 0x7D, 0x5E
 Replace 0x7D with 0x7D, 0x5D

4. Model of a Link

Rate R Mbps
Message
M bits Delay D seconds

 Abstract model is typically all we will need

 Other parameters that are important:


 The kind and frequency of errors
 Whether the media is broadcast or not
Message Latency
 How long does it take to send a message?

Message M

Delay D, Rate R
 Two terms:
 Propagation delay = distance / speed of light in media
• How quickly a message travels over the wire
 Transmission delay = message (bits) / rate (bps)
• How quickly you can inject the message onto the wire
 Later we will see queuing delay …

Relationships

 Latency = Propagation + Transmit + Queue


 Propagation Delay = Distance/SpeedOfLight
 Transmit Time = MessageSize/Bandwidth
One-way Latency

Dialup with a modem:


 D = 10ms, R = 56Kbps, M = 1024 bytes
 Latency = 10ms + (1024 x 8)/(56 x 1024) sec =
153ms!
Cross-country with T3 (45Mbps) line:
 D = 50ms, R = 45Mbps, M = 1024 bytes
 Latency = 50ms + (1024 x 8) / (45 x 1024*1024) sec
= 50ms!

 Either a slow link or long wire makes for large latency

Latency and RTT

 Latency is typically the one way delay over a link


 Arrival Time - Departure Time

Departure Arrival
Time Time

+ RTT

 The round trip time (RTT) is twice the one way delay
 Measure of how long to signal and get a response
Throughput
 Measure of system’s ability to “pump out” data
 NOT the same as bandwidth
 Throughput = Transfer Size / Transfer Time
 E.g., “I transferred 1000 bytes in 1 second on a 100Mb/s
link”
• BW?
• Throughput?
 Transfer Time = SUM OF
 Time to get started shipping the bits
 Time to ship the bits
 Time to get a response if necessary

Messages Occupy “Space” On the


Wire
 Consider a 1b/s network.

 Suppose latency is 16 seconds.


 How many bits can the network “store”
 This is the BANDWIDTH-DELAY product
 Measure of “data in flight.”
 1b/s * 16s = 16b

 Tells us how much data can be sent before a receiver sees any of it.
 Twice B.D. tells us how much data we could send before hearing
back from the receiver something related to the first bit sent.
 What are the implications of high B.D.?
A More Realistic Example

BD = 50ms * 45Mbps = 2.25 * 10^6 = 280KB

101100…11…0010101010101010101

Key Concepts
 We typically model links in terms of bandwidth and delay,
from which we can calculate message latency
 Different media have different properties that affect their
performance as links
 We need to encode bits into signals so that we can recover
them at the other end of the channel.
 Framing allows complete messages to be recovered at the far
end of the link

You might also like