CPS221 Lecture: Introduction to Net-Centric Computing

last revised 9/14/11

Objectives

1. To overview the history of computer networks in general and the Internet in

particular
2. To overview the major network standards bodies
3. To introduce the notion of a protocol
4. To overview the Internet as it exists today

Materials:
1. Projectable of Figure 1.2 from Kurose Ross
2. Projectable of RFC 2026
3. Projectable of Figure 1.2 from Forouzan
4. Projectable of Figure 1.1 from Forouzan

I. Introduction

A. The earliest computer systems functioned in isolation from one

another; but today it is almost impossible to conceive of a computer
system that is not networked in some way to other systems. This
includes not only desktop and laptop computers, but also devices such
as cell phones, entertainment devices, etc.

B. In common parlance, we regard “network” and “the Internet” as

virtual synonyms. But actually the term “the Internet” originates from
the term “internet” (lower-case I), which is simply a network that
connects other networks. There were (lowercase i) internets long
before the (uppercase I) Internet as we know it today became
dominant in the 1990’s.

1
C. Today, we want to look briefly at how this reality came about.

1. The history leading up to the Internet as we know it.

2. How the Internet copes with diversity.

If you think about it for a moment, the diversity in the Internet is

almost incredible.

a) There are a wide variety of types of devices connected to the

Internet
(1) Different brands of servers, desktop and laptop computers
(2) Many other kinds of devices.
(a) Cell phones
(b) Entertainment devices such as game consoles, digital
VCRs, etc
(c) As a rather extreme example, there is even an Internet
toaster that obtains a weather forecast from a
meteorological server and then burns an image of the
forecast on your morning toast. (Cited by Kurose/Ross p.
10)

b) These devices communicate with each other over a wide variety

of types of transmission media (cable systems, phone lines,
wireless ...)

c) These devices cooperate to accomplish a wide variety of tasks

web access, email, database access ...

D. We will also give a quick overview of the “Internet world” as it exists

today. In the process, we will look at the differing service
requirements of various kinds of Internet applications.

2
II. The History of the Internet

A. As I indicated earlier, the idea of a computer network and the idea of

an internet actually predate what we know today as the Internet.

B. The earliest computer networks

1. Often set up to connect computers within a single organization

(business)
a) Could be classified as LANs (local area networks) or WANs
(wide area networks) - e.g. a corporation with multiple sites
might use a WAN to support communication between
employees at different locations.

b) Supported services like email, information sharing.

Example: Gordon’s earliest network provided email between
users at Gordon, using usernames that didn’t require an “@”
symbol - e.g. my email address was simply “BJORK” at one
time.
c) Generally used proprietary technologies -

(1) e.g. Gordon’s earliest network used a product known as

“DECNET”, which was a product sold by Digital Equipment
Corporation - the manufacturer of our main computer
systems.

(2) IBM had a proprietary technology known as SNA (Systems

Network Architecture)

(3) Apple had a proprietary technology known as Appletalk

...

3
 2. There were also various dialup networks that supported services
like email and newsgroups (e.g. compuserve, genie, prodigy,
aol ...)

C. In 1967, the Advanced Research Projects Agency of the Department

of Defense (ARPA) proposed the creation of a network to connect
researchers working under defense contracts. By 1969, its first
implementation connected four universities: UCLA, UCSB, SRI, and
the University of Utah.

1. The ARPANET was the first network to use a technique known as

“packet switching”. Because this approach is pervasive in
networks today, it is useful to compare it to another approach
known as circuit switching.

a) The telephone network is an example of a circuit switched

network. In such a network

(1) A connection must be set up before communication can take

place (e.g. by dialing a phone number at one end and
answering at the other end)

(2) Once the connection is established information sent at one

end of the connection will be received at the other end
without any further identification of the intended recipient.

(3) When communication is done, it is necessary to explicitly

terminate the connection (e.g. by hanging up).
b) Packet switching resembles the approach used with postal mail:
information is sent over the network in discrete packets.

(1) Each packet carries with it the address of its destination.

4
 (2) Different packets between two communicating systems may
actually travel over different routes, and two packets sent at
about the same time may arrive in a different order from the
order in which they were sent.

(3) Short messages can be sent as a single packet; but longer

messages may need to be broken up into several smaller
packets.

c) A key difference between the two approaches is that with circuit

switching system resources remain dedicated to a particular
connection during the entire time it is in existence, whereas
with packet switching system resources are used on a per packet
basis.

This is why voice telephone service is charged on a “per

minute” basis, while postal mail charges per letter and text
messaging (which uses packet switching) charges per message.

2. The ARPANET (as it came to be called) grew to over 100 nodes,

but was always limited to defense contractor. However, it is
considered the major precursor of today’s Internet because many of
the key technologies used by the Internet (e.g. TCP/IP) were first
developed there.

3. A number of other similar networks arose to support non-defense

research, such as the NSFNET and CSNET.

4. What eventually emerged in the early 1990’s was the concept of

the Internet as a network of networks. What we know today as the
Internet is actually the successor to these earlier networks.

D. Probably no single development contributed as much to the

popularization of the Internet as the development of the World-Wide
Web.

5
 1. This was originally developed by Tim Berners-Lee in 1989 at
CERN (the European Organization for Nuclear Research located in
Switzerland.)

2. A key turning point was the development of the Mosaic web

browser at the National Center for Supercomputers (NCSA) in
1993. (Many of the original developers of Mosaic worked on a
browser known as Netscape which became the precursor of the
Mozilla project and Firefox)

III.How Diverse Systems can Communicate: Protocols

A. So how do diverse systems communicate with one another?

They do so using protocols.

B. A protocol is simply a set of rules that facilitate communication.

Example: PROJECT Figure 1.2 from Kurose/Ross

C. In the world of computer networks, there are protocols at multiple

levels

1. There are physical protocols, that specify how systems represent

binary information. For two devices to connect to each other, they
must both represent binary information in the same way.

2. There are protocols that govern how messages are routed over the
network..
Example from postal mail: the destination address goes near the
center of the envelope, and the return address goes in the upper-
left-hand corner. Imagine what would happen if you switched the
two on a letter you sent!

6
 3. There are protocols that govern how packets are interpreted by
various applications.

a) Web requests and pages follow a protocol known as http

(hypertext transfer protocol). (That’s what the http in a URL
stands for. Note that there is a different protocol (https) used
for encrypted web operations)

b) Email messages follow a protocol known as smtp (simple mail

transfer protocol)
...

IV.How Protocols are Agreed Upon: Standards

A. Of course, protocols are only useful if both parties agree to use the
same protocol.

B. Standards are of two types:

1. De facto standards are widely used, though they may originate

from a single source.

Example: Microsoft Word format is a defacto standard for word-

processing, widely accepted because Word is so commonly used.

2. De jure standards are formally promulgated by a standards body,

typically after a process that involves opportunity for comment
from the parties most affected by the standard.

C. In the Internet world, commonly-used standards come from a variety

of standards bodies - many of which deal with many areas, not just the
Internet.

1. Standards bodies that deal with a broad range of standards (not just
electronics)

7
 a) ISO - International Standards Organization. It is an
international body whose membership is drawn mainly from
various national standards bodies.

(1) In the world of networks, ISO standardized a reference

model for network protocol hierarchies which we will talk
about later (the ISO Open Systems Interconnect Model).

(2) In the world of computers, ISO is also responsible for a

standard format for CDs and has created standards for some
programming languages.

b) ANSI (American National Standards Institute) is the US

affiliate of ISO.

(1) In the world of computers, it is the body that has

standardized the ASCII code that is commonly used for the
English alphabet.

(2) ANSI has also standardized a number of programming

languages.

2. Standards bodies that deal primarily with electronics broadly (not

limited to telecommunications).

a) IEEE - the Institute of Electrical and Electronics Engineers.

b) EIA - Electronics Industry Association

3. Standard bodies that deal primarily with telecommunications and

Internet-related standards

8
 a) ITU-T - The International Telecommunications Union, and its
committee CCITT - the Consultive Committee for International
Telegraphy and Telephony.

b) W3C - The World-Wide-Web Consortium.

D. In the Internet world, standards go through a process that is managed

by the Internet Society through its Internet Activities Board (IAB) and
the Internet Engineering Task Force (IETF).

1. The standardization process is built on documents known as

“Requests for Comment” (RFC). (The name reflects the history of
this process).

2. The process as it exists today is itself defined by an RFC:

PROJECT RFC 2026. (Note references to updates that pertain to

specific portions)

3. Not all RFC’s pertain to standards. (In fact, there is a tradition of a

joke RFC being published each year on April 1; an obituary of
John Postel, the original editor of the RFC series, is RFC 2468!)

4. Any RFC can be accessed by a URL of the form http://

tools.ietf.org/html/rfcXXXX

(Demo accessing RFC2468 this way)

5. When an RFC pertains to a standard, it goes through a process

(known as a sequence of maturity levels) that may - or may not -
result in it becoming a standard

PROJECT Forouzan figure 1.2.

9
V. The World of the Internet

A. The Internet - as it exists today - is a network of networks that can be

depicted as follows:

PROJECT Forouzan figure 1.1

B. Nodes on the Internet are identified in one of two ways.

1. Each node has a unique address known as an IP number.

a) Most IP addresses are based on a standard known as ipv4,
which uses 4 bytes. (However, v4 refers to this being version 4
of the protocol - not the fact that it uses 4 bytes!)
(1) IP addresses are “owned” by ISPs and are assigned by
higher level ISPs to lower level ISPs and local users. IP
addresses are used to actually route traffic over the network.
(2) The most common way to write an IP address is as dotted
octets. (The term octet comes from the fact that each
number is the decimal equivalent of 8 bits)
SHOW Network pane in System Preferences - my IP
(3) While this potentially provides 4 billion addresses, not all
addresses are actually used.
(a) Some are reserved for special purposes.
(b) An IP number is actually regarded as being divided into
two parts - a network number and the IP number of a
specific host on that network. (Recall that the Internet is
a network of networks).
SHOW how this is specified by the subnet mask for my
computer.

10
 (4) Moreover, even 4 billion addresses are actually not enough
for all of the hosts connected to the internet (not just
computers, but smart phones, applicances ...)

(5) The usefulness of ipv4 has been extended by using a

technique known as Network Address Translation (NAT). I
will use the specific way this works out at Gordon as an
example:
(a) Gordon actually owns only 70 IP addresses
(216.236.251.129 .. 198)
(b) Machines on campus have IP numbers of the form
172.XXX.YYY.ZZZ or 192.168.YYY.ZZZ or
10.10.XXX.YYY - all in ranges of IP numbers reserved
by ipv4 for private networks.
(c) When a machine on campus wants to contact an off-
campus site, Gordon’s gateway replaces its “on campus”
IP number with one of the external IPs Gordon owns.
When a response comes back, Gordon’s gateway replaces
the “external” IP with the “internal” IP number and sends
it to the originating system.
(d) Gordon also permanently assigns several of its external
IPs to systems on campus (e.g. an incoming message on
216.236.251.139 is sent to the CS web server, known
by the internal IP address 172.16.50.17 and by the
URL www.cps.gordon.edu.)

b) One will sometimes see a range of IP addresses denoted by a

notation such as 172.16.0.0/12. This specifies a range in which
the upper 12 bits correspond to 172.16.0.0 but the remaining 20
bits can take on any value - so the range goes from

10101010 00010000 00000000 00000000 (172.16.0.0) to

10101010 00011111 11111111 11111111 (172.31.255.255)

11
 c) There is a newer standard - called ipv6 - that uses 128 bit IP
numbers. While this standard was defined in 1998, it is only
slowly being implemented.

2. Though the IP address of a node is what is actually used in Internet

routing, hosts are more often identified by a URL (e.g.
www.gordon.edu or go.gordon.edu).

a) A full URL typically consists of three or four parts.

Example: When this lecture is posted online, its URL wlll be

http://www.cs.gordon.edu:80/courses/cps221/
lectures-2011/Introduction%20%to%20Net-Centric
%20Computing.pdf

http: specifies the protocol

www.cs.gordon.edu specifies the host machine
80 specifies the port number (generally omitted since often
implied by the protocol - e.g 80 is the port normally used for
http)
courses/cps221/lectures-2011/Introduction%20%to
%20Net-Centric%20Computing.pdf specifies a path on the
host machine

b) To actually route traffic to a system identified by a URL, the

host name portion of the URL must be translated into an IP
number. This is handled by a system designated as a name
server, which maintains a table of URLs and the corresponding
IP numbers.

(1) To prevent the name tables from becoming very cluttered, a

hierarchical scheme based on domains is actually used.
Each domain has one or more domain name servers (DNS)

12
 (a) There is a DNS for the edu domain, which provides the
IP number for a DNS for the gordon.edu domain.

(b) There is a DNS for the gordon.edu domain, which

provides the IP number for a DNS for the
cs.gordon.edu domain.

(c) There is a DNS for the cs.gordon.edu domain, which

actually provides the IP number for the
www.cs.gordon.edu host.

(2) While Gordon is responsible for all URLs within the

gordon.edu domain (including the cs subdomain), top-
level domains are managed by the Internet Corporation for
Assigned Names and Numbers (ICANN) through a large
number of domain name registrars that are allowed to
register names in various domains.

c) Note that, while the host name portion of a URL must be

mapped to a specific IP number, several different host name
portions of a URL might all refer to the same IP - e.g.

files.cs.gordon.edu and joshua.cs.gordon.edu both

refer to the same machine (172.16.50.16)

C. Of course, there are a huge variety of applications that utilize the

Internet. One way of categorizing them is in terms of their service
requirements.

1. Bandwidth - the rate of information transfer needed - often

specified in units such as MB / second.

13
2. Tolerance for end-to-end delay (the interval between the time the
information is sent and the time it is received.)

3. Tolerance for loss of information in transit (dropped packets).

4. Examples:

a) Electronic mail requires relatively small bandwidth; has high

tolerance for delay, but low tolerance for information loss.
(In fact, in contrast with many other applications where larger-
size units of information are transmitted, the loss of a packet
could result in the lost of an entire email with no indication to
either the sender or the receiver that this has occurred.)

b) The bandwidth requirements for web page access varies

depending on the page contents; there is moderate tolerance for
delay (up to a few seconds), and some information loss can be
tolerated (the requester can request the page again; if a portion
of a page is lost the recipient's browser will typically detect this
and request a re-send of the lost portion.)

c) For streaming multimedia the bandwidth requirements tend to

be much higher and the tolerance for delay tends to be smaller,
though some loss of information can be tolerated.

d) For interactive applications (e.g. Skype), high bandwidth is

again needed, tolerance for delay is very small, though again
some loss of information can be tolerated.

14