Chapter 2: Network architecture and

performance
 Circuit switching and packet switching
 Layers and functions of layers - OSI and TCP/IP model

 Name and address

 Principles of Internet Design

 The factors of network performance

Network architecture = topo +
protocol
network edge:
applications and
hosts
 access networks,
physical media:
wired, wireless
communication
 links
network core:
 interconnected
routers
 network of
networks
 The network edge
 end systems (hosts):
 run application programs
 e.g. Web, email
 at “edge of network” peer-peer
 client/server model
client host requests,

receives service from
always-on server client/server
 e.g. Web browser/server;
 peer-peer model:
email client/server
 minimal (or no) use of
dedicated servers
 e.g. Skype, BitTorrent
Company access: local area networks

 company/univ local area

network (LAN) connects
end system to edge
router
 Ethernet:
 10 Mbs, 100Mbps,
1Gbps, 10Gbps
Ethernet
 modern
configuration: end
systems connect into
Ethernet switch
Wireless access networks

 shared wireless access

network connects end router
system to router
 via base station aka “access base
point” station
 wireless LANs:
 802.11b/g (WiFi): 11 or 54
Mbps
 wider-area wireless access mobile
 provided by telco operator hosts
 ~1Mbps over cellular system
(EVDO, HSDPA)
 next up (?): WiMAX (10’s Mbps)
Home networks

Typical home network components:

 DSL or cable modem

 router/firewall/NAT

 Ethernet

 wireless access point

wireless
to/from laptops
cable router/
cable
modem firewall
headend
wireless
access
Ethernet point
The Network Core
 mesh of interconnected
routers
 the fundamental
question: how is data
transferred through
net?
 circuit switching:
dedicated circuit
per call: telephone
net
 packet-switching:
data sent thru net
in discrete “chunks”
Network Core: Circuit Switching

End-end resources reserved for “call”

 link bandwidth, switch capacity
 dedicated resources: no sharing
 circuit-like (guaranteed) performance
 call setup required
Network Core: Circuit Switching
network resources  dividing link
(e.g., bandwidth) bandwidth into
divided into “pieces”
 frequency division
“pieces”
 time division
 pieces allocated to
calls
 resource piece idle if
not used by owning
call (no sharing)
Circuit Switching (Chuyển mạch kênh)

 Tài nguyên được gán cho mỗi kênh

 Không chia sẻ cho người khác ngay cả khi kênh đang rỗi
Network Core: Packet Switching

each end-end data stream resource contention:

divided into packets  aggregate resource
 user A, B packets share demand can exceed
network resources amount available
 each packet uses full  congestion: packets
link bandwidth queue, wait for link
 resources used as use
needed  store and forward:
packets move one
Bandwidth division into
“pieces”
hop at a time
 Node receives complete
Dedicated allocation packet before forwarding
Resource reservation
 Packet Switching: Statistical
Multiplexing
100 Mb/s
A Ethernet statistical multiplexing C

1.5 Mb/s
B
queue of packets
waiting for output
link

D E

Sequence of A & B packets does not have fixed

pattern, bandwidth shared on demand  statistical
multiplexing.
TDM: each host gets same slot in revolving TDM frame.
Packet-switching: store-and-
forward
L
R R R

 takes L/R seconds to Example:

transmit (push out)  L = 7.5 Mbits
packet of L bits on to  R = 1.5 Mbps
link at R bps
 transmission delay =
 store and forward:
entire packet must 15 sec
arrive at router before
it can be transmitted
on next link
 delay = 3L/R more on delay shortly …
(assuming zero
propagation delay)
Packet Switching (Chuyển mạch gói)

 Toàn bộ băng thông được chia sẻ cho tất cả mọi người

 Nếu còn băng thông thì bất kỳ ai cũng có thể sử dụng
 Packet switching versus circuit
switching
Packet switching allows more users to use
 network!
1 Mb/s link
 each user:
 100 kb/s when “active”
 active 10% of time
(0.1) N users
1 Mbps link
 circuit-switching:
 10 users P(n,N) =(N!) ( p^n) (1-p)^(N-n)
(n!) (N-n)!
 packet switching:
P(11,35) =(35!) ( 0.1^11) (0.9^24)
with 35 users,
probability > 10 active
(11!) (24!)
at same time is less = 0.0003
Q: how did we get value 0.0004?
than .0004 P(12,35) = 0.0001
 Prob. Host is Active: 0.1 No. Hosts 35

Number Active (n) Prob(n,N) Prob( >n,N) Poisson(n,P*N)

0 0.0250 0.9750 0.0302 0.9608
1 0.0973 0.8776 0.1057 0.8641
2 0.1839 0.6937 0.1850 0.6792
3 0.2247 0.4690 0.2158 0.4634
4 0.1998 0.2693 0.1888 0.2746
5 0.1376 0.1316 0.1322 0.1424
6 0.0765 0.0552 0.0771 0.0653
7 0.0352 0.0200 0.0385 0.0267
8 0.0137 0.0063 0.0169 0.0099
9 0.0046 0.0017 0.0066 0.0033
10 0.0013 0.0004 0.0023 0.0010
11 0.0003 0.0001 0.0007 0.0003
12 0.0001 0.0000 0.0002 0.0001
13 0.0000 0.0000 0.0001 0.0000
14 0.0000 0.0000 0.0000 0.0000
15 0.0000 0.0000 0.0000 0.0000

Exact: Prob(n,N) = N! Pn (1-P)N-n

n! (N-n)!

Approximation: Poisson(n,M) = exp(-M)Mn/n!

where the mean M= P*N = 3.5 hosts (P=10%=0.1)Introduction
Packet switching versus circuit
switching
Is packet switching a winner?
 great for bursty data

 resource sharing

 simpler, no call setup

 excessive congestion: packet delay and loss

 protocols needed for reliable data transfer,

congestion control
 Application
Circuit switching: PSTN (Published Switched
Telephone Network), ISDN
Packet switching: ATM, DSL, MPLS, Satellite,
Wireless
How do the loss and delay
occur?
packets queue in router buffers
 packet arrival rate to link exceeds output link
capacity
 packets queue, wait for turn
packet being transmitted (delay)

B
packets queueing (delay)
free (available) buffers: arriving packets
dropped (loss) if no free buffers
Four sources of packet
delay
 1. nodal processing:  2. queueing
 check bit errors  time waiting at output

 determine output link link for transmission

 depends on
congestion level of
router

transmission
A propagation

B
nodal
processing queueing
Delay in packet-switched
networks
3. Transmission delay: 4. Propagation delay:
 R=link bandwidth  d = length of physical

(bps) link
 L=packet length  s = propagation speed

(bits) in medium (~2x108

 time to send bits into m/sec)
link = L/R Note:
 s and R are
propagation very
delay =
transmission
different quantities!
d/s
A propagation

B
nodal
processing queueing
 Caravan analogy

100 100
km km
ten-car toll toll
caravan booth booth
 cars “propagate” at  Time to “push” entire
100 km/hr caravan through toll
 toll booth takes 12 sec to booth onto highway =
service car (transmission 12*10 = 120 sec
time)  Time for last car to
 car~bit; caravan ~ packet propagate from 1st to
 Q: How long until caravan 2nd toll both:
is lined up before 2nd toll 100km/(100km/hr)= 1 hr
booth?  A: 62 minutes
Caravan analogy (more)

100 100
km km
ten-car toll toll
caravan booth booth

 Cars now “propagate”

 Yes! After 7 min, 1st car
at at 2nd booth and 3 cars
1000 km/hr still at 1st booth.
 Toll booth now takes 1
 1st bit of packet can
min to service a car arrive at 2nd router
before packet is fully
 Q: Will cars arrive to transmitted at 1st router!
2nd booth before all  See Ethernet applet at AWL
cars serviced at 1st Web site
booth?
Nodal delay

d nodal d proc  d queue  d trans  d prop

 dproc = processing delay
 typically a few microsecs or less
 dqueue = queuing delay
 depends on congestion
 dtrans = transmission delay
 = L/R, significant for low-speed links
 dprop = propagation delay
 a few microsecs to hundreds of msecs
“Real” Internet delays and routes

 What do “real” Internet delay & loss look like?

 Traceroute program: provides delay
measurement from source to router along end-
end Internet path towards destination. For all
i:
 sends three packets that will reach router i on path
towards destination
 router i will return packets to sender
 sender times interval between transmission and reply.
3 probes 3 probes

3 probes
“Real” Internet delays and routes

tracert: gaia.cs.umass.edu to www.eurecom.fr

Three delay measurements from
gaia.cs.umass.edu to cs-
1 cs-gw (128.119.240.254) 1 ms 1 ms 2 ms
gw.cs.umass.edu
2 border1-rt-fa5-1-0.gw.umass.edu (128.119.3.145) 1 ms 1 ms 2 ms
3 cht-vbns.gw.umass.edu (128.119.3.130) 6 ms 5 ms 5 ms
4 jn1-at1-0-0-19.wor.vbns.net (204.147.132.129) 16 ms 11 ms 13 ms
5 jn1-so7-0-0-0.wae.vbns.net (204.147.136.136) 21 ms 18 ms 18 ms
6 abilene-vbns.abilene.ucaid.edu (198.32.11.9) 22 ms 18 ms 22 ms
7 nycm-wash.abilene.ucaid.edu (198.32.8.46) 22 ms 22 ms 22 ms trans-oceanic
8 62.40.103.253 (62.40.103.253) 104 ms 109 ms 106 ms
9 de2-1.de1.de.geant.net (62.40.96.129) 109 ms 102 ms 104 ms link
10 de.fr1.fr.geant.net (62.40.96.50) 113 ms 121 ms 114 ms
11 renater-gw.fr1.fr.geant.net (62.40.103.54) 112 ms 114 ms 112 ms
12 nio-n2.cssi.renater.fr (193.51.206.13) 111 ms 114 ms 116 ms
13 nice.cssi.renater.fr (195.220.98.102) 123 ms 125 ms 124 ms
14 r3t2-nice.cssi.renater.fr (195.220.98.110) 126 ms 126 ms 124 ms
15 eurecom-valbonne.r3t2.ft.net (193.48.50.54) 135 ms 128 ms 133 ms
16 194.214.211.25 (194.214.211.25) 126 ms 128 ms 126 ms
17 * * *
18 * * * * means no response (probe lost, router not
19 fantasia.eurecom.frreplying)
(193.55.113.142) 132 ms 128 ms 136 ms
 Packet loss
 queue (aka buffer) preceding link in buffer
has finite capacity
 packet arriving to full queue dropped (aka
lost)
 lost packet may be retransmitted by
previous node,buffer
by sourcepacket
end
being
system, or
transmitted
(waiting area)
A
not at all

B
packet arriving to
full buffer is lost
Exercise
L
R1 R2 R3

 Consider a packet of length L which begins at end system A and

travels over three links to a destination end system. These three
links are connected by two packet switches. Let di, si, and Ri
denote the length, propagation speed, and the transmission rate
of link i, for i = 1, 2, 3. The packet switch delays each packet by
dproc. Assuming no queuing delays, in terms of di, si, Ri,(i =
1,2,3), and L, what is the total end-to-end delay for the packet?
 Suppose now the packet is 1,500 bytes, the propagation speed
on all three links is 2.5·108 m/s, the transmission rates of all
three links are 2 Mbps, the packet switch processing delay is 3
msec, the length of the first link is 5,000 km, the length of the
second link is 4,000 km, and the length of the last link is 1,000
km. For these values, what is the end-to-end delay?
 Throughput
 throughput: rate (bits/time unit) at
which bits transferred between
sender/receiver
 instantaneous: rate at given point in time
 average: rate over longer period of time

server,
server sends linkthat
with pipe capacity
can carry link that
pipe capacity
can carry
filebits
of F bits fluid at rate
Rs bits/sec Rfluid at rate
c bits/sec
to send
(fluid) to pipe
into client R bits/sec) Rc bits/sec)
s
Throughput vs bandwidth
 Bandwidth is the maximum amount
of data that can be transferred through
a network for a specified period of
time.
 Throughput is the actual amount of
data that can be transferred through a
network during a specified time period
 Throughput (more)
 Rs < Rc What is average end-end
throughput?
Rs bits/sec Rc bits/sec

 Rs > Rc What is average end-end

throughput?
Rs bits/sec Rc bits/sec

bottleneck link
link on end-end path that constrains end-end
throughput
Throughput: Internet
scenario
Rs
 per-connection Rs Rs
end-end
throughput:
R
min(Rc,Rs,R/10)
Rc Rc
 in practice: Rc or
Rs is often Rc

bottleneck
10 connections (fairly) share
backbone bottleneck link R
bits/sec
Protocol “Layers”
Networks are
complex!
 many “pieces”: Question:
Is there any hope of
 hosts organizing structure of
 routers network?
 links of various
media Or at least our discussion
of networks?
 applications
 protocols
 hardware,
software
Organization of air travel

ticket (purchase) ticket (complain)

baggage (check) baggage (claim)

gates (load) gates (unload)

runway takeoff runway landing

airplane routing airplane routing

airplane routing

 a series of steps
Layering of airline
functionality

ticket (purchase) ticket (complain) ticket

baggage (check) baggage (claim baggage

gates (load) gates (unload) gate

runway (takeoff) runway (land) takeoff/landing

airplane routing airplane routing airplane routing airplane routing airplane routing

departure intermediate air-traffic arrival

airport control centers airport

Layers: each layer implements a service

 via its own internal-layer actions
 relying on services provided by layer
below
Why layering?

Dealing with complex systems:

 explicit structure allows identification,
relationship of complex system’s pieces
 layered reference model for discussion
 modularization eases maintenance, updating
of system
 change of implementation of layer’s
service transparent to rest of system
 e.g., change in gate procedure doesn’t
affect rest of system
 layering considered harmful?
 OSI – Open System Interconnection

Layer 7

Layer 6

Layer 5

Layer 4

Layer 3

Layer 2

Layer 1
Internet protocol stack
 application: supporting network
applications application
 FTP, SMTP, HTTP
 transport: process-process data transport
transfer
 TCP, UDP
network
 network: routing of datagrams (packet)
from source to destination (host-to-host
communication).
link
 IP, routing protocols
 link: data transfer between neighboring physical
network elements
 PPP, Ethernet
 physical: bits “on the wire”
OSI reference model
 presentation: allow applications
to interpret meaning of data, e.g., application
encryption, compression, presentatio
machine-specific conventions n
 session: synchronization, session
checkpointing, recovery of data transport
exchange
network
 Internet stack “missing” these
layers! link
 these services, if needed, physical
must be implemented in
application
 needed?
 OSI reference model
1-39
OSI vs. TCP/IP
1-41
Internet Layer Model

 Example: Data transmission from source, through

intermediate node (router), to destination
PDU: Đơn vị dữ liệu giao
thức
SAP: Điểm truy cập dịch vụ
Encapsulation and
Decapsulation

 Encapsulation (Sender): Each layer adds control information to

the header and transmits it to the lower layer.
 Decapsulation (Receiver): Each layer processes the packet
based on the information in the header, then removes the
header, taking the data and forwarding it to the upper layer.
Encapsulation and Decapsulation
 source Encapsulatio
message
segment Ht
M
M
applicatio
n
n
1-47
datagram Hn Ht M transport
frame Hl Hn Ht M network
link
physical link
physical

switch

destination Hn H t M network
M applicatio
H l Hn H t M link Hn H t M
Ht M n physical
Hn H t M transport
H l Hn H t M network router
link
physical
Ưu điểm của kiến trúc phân tầng

 Chia nhỏ cho phép xác định dễ dàng chức năng mỗi tầng
 Các tầng hoạt động độc lập
 Tầng trên chỉ quan tâm đến việc sử dụng tầng dưới

mà không quan tâm đến các tầng xa hơn

 Cho phép định nghĩa giao diện chung giữa các tầng

 Khả năng mở rộng

 Mềm dẻo, linh hoạt với các công nghệ mới
 Trao đổi giữa các tầng đồng mức

 Có thể cải tiến hệ thống bằng cách thay thế một công

nghệ mới của tầng tương ứng: ISDNADSLFTTH;

IPv4IPv6
 Nếu không phân tầng
 Khi muốn thay đổi, phải làm toàn bộ…
Identifiers on the Internet
MAC address

 Physical address/MAC address

 Used at Data link layer

 Fixed on NIC network card

 Used to address computers in broadcast networks

IP address
 Used in IP protocol (network layer)
 The value depends on each network,
each network card is assigned an IP
address
 Used to identify a computer in the
network
 For example:
 130.145.215.9 (ipv4)
 2001:200:0:8803::53 (ipv6)
Port number

 Address used in transport layer

 A secondary index, used with IP address

 Applications are identified by an IP

address and a port number

 Similar to the number of rooms in a

house
 Address: 11 Nguyen Dinh Chieu, District 1,
HCMC => IP Address
 Room 12 => Port number
Domain Name
 Domain Name
 FQDN: Fully Qualified Domain Name
 Là tên của một máy tính hay của một mạng
máy tính, sử dụng tên (chữ cái, chữ số)
 www.uef.edu.vn
 www.vnexpress.net
 www.moet.gov.vn
Không gian tên miền
Tên và địa chỉ
 Trước khi truyền tin, máy trạm phải được xác định:
 Bởi một địa chỉ IP, hoặc
 Bởi một tên miền (thuận tiện cho NSD)
 Tên:
 Độ dài thay đổi
 Dễ nhớ cho con người
 Không liên quan tới vị trí vật lý của máy
 Địa chỉ:
 Độ dài cố định
 Dễ cho máy tính để xử lý
 Liên quan tói vấn đề chọn đường
Chuyển đổi địa chỉ

 Cơ chế cho phép tìm một địa chỉ IP từ một tên miền và ngược lại
 Máy chủ tên miền (DNS):
 Một phương thức được sử dụng để chuyển đổi địa chỉ
 Được sử dụng rộng rãi trên Internet
Chuyển đổi địa chỉ - ví dụ
Công cụ nslookup
 Nhập tham số www.hedspi.hut.edu.vn
 Chuyển đổi “Tên  Địa chỉ” được thực
hiện
 Lệnh nslookup
 nslookup www.hedspi.hut.edu.vn
ARP: Chuyển đổi địa chỉ vật lý – địa chỉ
IP
 Address Resolution Protocol
 Dùng để tìm địa chỉ vật lý từ địa chỉ IP

 Ví dụ ARP table
Exercise
 Advantage and Disadvantage of OSI
model and TCP/IP stack?
 Reference: