EECS 563 Fall 2013 Review Test 2 - KU ITTCMicrosoft PowerPoint - Review-Test-2-563-Fall-2013.pptx...
Transcript of EECS 563 Fall 2013 Review Test 2 - KU ITTCMicrosoft PowerPoint - Review-Test-2-563-Fall-2013.pptx...
1Review Test 2
EECS 563 Fall 2013
Review Test 2
2Review Test 2
Network traffic � Request for resources
� Rate of requests= λ � Average resource hold time = Th
� Common Assumptions– Time between arrivals ~ exponentially– Holding time ~ exponentially
� Voice� λ = call/sec� Th= average holding time sec/call� Load = λ*Th Erlangs
3Review Test 2
Network traffic � Packet Voice
�Constraint� limit on end-to-end delay� Packetization time (ms/packet)� Role of the Jitter buffer
– Too small � lose packets– Too large � increase delay
4Review Test 2
Network traffic � Packet Voice
�Components of end-to-end delay (delay budget) & trade-offs– Packetization time– Clocking time– Network delay (random)– Jitter buffer
�VoIP Rate calculation
5Review Test 2
Network traffic � Video
�MPEG� Frame types
� Data� Burstyness = (Peak rate)/(Average rate)� Large Burstyness leads to effective statistical multiplexing.
6Review Test 2
Performance Anaysis� Given
� Traffic
�QoS
� Design system
� Find C
� System size
� Find Number of Servers� Assumptions
� Service time ~ Exponentially distributed
� Interarrival time ~ Exponentially distributed
7Review Test 2
M/M/1
1[ ]
1
L
CE Tµ λ ρ
= =− −
Where Average Service Rate, Average Arrival Rate, C= Capacity, L = Average Packet Lengthµ λ= =
inrL
C C
λρ λ
µ= = =
[ ]
[ ] 1
P System Busy
P System Empty
ρ
ρ
=
= −
Given on Test
8Review Test 2
M/M/1/N and M/M/N/N
� M/M/1/N
� M/M/N/N � Erlang B
1
(1 )
1
N
B NP
ρ ρρ +
−=
− Given on Test
Table Given on Test
9Review Test 2
MAC
� Scaling & trade-offs WRT: rate (b/s), number of users, and size (km)
� Deterministic (Polling)
�Operation (why called deterministic)
�Calculate effective rate & efficiency'
' where
As Max
a Ring LatencyL
C
a S
ττ= =
↑ ↓
10Review Test 2
MAC
� Random Access
�Collision process
– Time vulnerable to collision
– Detecting Collisions
� Time
– Unslotted
– Slotted
� Role of backoff
11Review Test 2
MAC (Random Access-continued)
� Types (all can be slotted/unslotted)� ALOHA (for unslotted Smax=18%, for slotted Smax= 36%)
� CSMA
– p-persistent (1-persistent)
– Non-persistent
– CSMA/CD
� Leads to specification of Min/Max Packet size
where End-to-End Propagation TimeaL
C
ττ= =
As and as 1,Max Max
a S a S ALOHA↑ ↓ → →
12Review Test 2
MAC
� Collision Free Protocols
� Random Access and Reservation Systems
� In upstream send requests to transmit – Use part of frame (contention slots) to send requests
– Use random access to share contention slots
� Receive grants in the downstream
� No contention in downstream
� If no grant in downstream then assume collision for the request, backoff and resend request in upstream
13Review Test 2
MAC- Ethernet
� IEEE 802.3
� Evolution
– Bus
– Hub
– Switch
– 10 Mb/s � 100 Gb/s
� Role of CSMA/CD
14Review Test 2
Network Elements
� Repeater
� Bridge
� Switch
� Router� Layer 2 Switch
� Layer 3 Switch
� Layer 4 Switch
� Layer “Any” Switch
15Review Test 2
MAC- Wireless Networks
� Issues
�Noise
� Signal Fading
�Hidden terminal
� RTS/CTS
16Review Test 2
MAC- Cable Networks
� DOCSIS
� Access protocol
� CM, Headend, CMTS
17Review Test 2
DLC
� Goal� point-to-point error free link
� Functions
� Framing� Flags & bit stuffing
� Error recovery
� Flow control
18Review Test 2
DLC
� Sliding window flow control
� n bits/SN in packet header
� Max window �N= 2n-1
� N=1 � Stop and Wait
� When to retransmit?– Timeout
– RNR (NACK)
� What to retransmit?– SN
– Go-back-N
– Selective Repeat
19Review Test 2
DLC
� Piggybacking
� Frame structure
� HDLC
20Review Test 2
DLC
� Performance
� Stop&Wait*
� Sliding window*
#
Time to tx #
eff
eff
R
R
bitsR
bits
η =
=
&
1
21
stop wait
f
R
n
ητ
=+
21 if N 1
2 if 1
21
f
SlidingWindow
f
f
R
n
N RN
R n
n
τ
η ττ
≥ +
= < +
+
* Understand assumptions behind these equations
21Review Test 2
DLC
� Open Loop Control
�DE bit
�Methods
– CIR, Bc, Be
– Token bucket� Average rate
� Maximum burst size