Networks

1
Networks
Network Theory and other stuff

• The Internet may be described as part technology
  and part human interaction. To describe it as
  one or the other is not quite accurate. Unlike
  other technologies, it does not do anything in
  the absence of the human mind in fact, the
  human mind is the sole source of its viability.
  Accordingly, the destiny of the Net will be
  determined by the interaction of two adaptive
  agents
• The systems and software of the Net
• Its human users

Source Valovic, Digital Mythologies
2
Networks
Network Theory and other stuff
The Web is a Network Not only that, the Web is
complex network so says Sir Tim Berners-Lee (and
just about every other scientist in the world who
is doing research on networks or complexity
theory).
So lets take this as a given. The Web is a
Complex Network
3
Graphic view of the Web (by tracing links)
Each color on this Opte map represents a region
North America, blue Europe/Middle East/Central
Asia/Africa, green Latin America, yellow Asia
Pacific, red Unknown, white. (Image
Opte.org) Source NewScientist.com
4
Network Theory and other stuff

• Elements of complex systems
• They are dynamic in the sense they are constantly
  changing.
• They are adaptive, which is to say they evolve to
  benefit themselves and to insure their survival.
• The adaptations are controlled to some extent by
  the interactions of the entities that comprise
  the system.
• The control is typically highly dispersed.
• They exhibit many levels of organisation.
• They are comprised of many niches.
• They are self-organising.
• New elements or entities emerge from complex
  systems.
• The emergent elements are not necessarily
  predictable from analysis of the individual parts
  of the system.
• Complex systems are defined by relationships
  between components more than by describing its
  constituent parts.

5
Network Theory and other stuff

• Research published by scientists at Notre Dame in
  1999 indicated that there were fundamental
  attributes of most networks, including the
  Internet and the Web, in that they
• Exhibited rapid and/or consistent growth.
• Exhibited a power law distribution.
• Exhibited forms of preferential attachment.

6
Network Theory and other stuff
Rapid and/or Consistent Growth
Netcraft's latest Web survey found 101,435,253
websites in November 2006. Not all of these sites
are live some are "parked" domains, while others
are abandoned weblogs that haven't been updated
in ages. But even if only half the sites are
maintained, there are still more than 100 M sites
that people pay to keep running.
Total sites across all domains
Source Jakob Nielson http//www.useit.com/alertbo
x/web-growth.html

• As the chart shows, the number of Websites has
  experienced three growth stages
• 1991-1997 Explosive growth, at a rate of 850
  per year.
• 1998-2001 Rapid growth, at a rate of 150 per
  year.
• 2002-2006 Maturing growth, at a rate of 25 per
  year.

7
Network Theory and other stuff
Rapid and/or Consistent Growth
The 100 million site milestone caps an
extraordinary year in which the Internet has
already added 27.4 million sites, easily topping
the previous full-year growth record of 17
million from 2005. The Internet has doubled in
size since May 2004, when the survey hit 50
million. Blogs and small business web sites have
driven the explosive growth this year, with huge
increases at free blogging services at Google and
Microsoft. Domain industry juggernauts Go Daddy
(U.S.) and 11 Internet (Germany) have also seen
strong demand for low-priced domain names and
shared hosting accounts.
Total sites across all domains
8
Network Theory and other stuff
Power Law Distribution
Measured by inbound links
WSU
9
Network Theory and other stuff
Power Law Distribution
What is a Normal Distribution?
10
Network Theory and other stuff
Power Law Distribution
What is a Normal Distribution?
A normal distribution of data means that most of
the examples in a set of data are close to the
"average," while relatively few examples tend to
one extreme or the other.
11
Network Theory and other stuff
Power Law Distribution
Until relatively recently (the mid 1990s) it was
generally assumed that many (most?) networks
exhibited a normal distribution of nodes and
edges.
A normal distribution of data means that most of
the examples in a set of data are close to the
"average," while relatively few examples tend to
one extreme or the other.
12
Network Theory and other stuff
Power Law Distribution
By 1999 several scientists had published papers
indicating the nodes of Web did NOT form a normal
distribution when measured by the edges (or
inbound links), but instead formed a Power
Law distribution.
Adamic and Huberman The Webs Hidden Order,
Communications of the ACM, 2001 Barabasi and
Albert Emergence of Scaling in Random
Networks, Science, Vol 256, Oct, 1999

• Power laws as related to websites may be verbally
  represented as
• a very few sites that rank very high in the
  number of inbound links
• a larger number of sites with close to median
  numbers of inbound links
• a great number of sites with very few inbound
  links.
• In short, the Web has many small elements, and
  few large ones. A few sites have millions of
  pages but millions of sites have only a few
  pages. A few sites have millions of inbound
  links, but millions of sites have only a few
  inbound links.

13
Network Theory and other stuff
Power Law Distribution

• Power laws as related to websites may be verbally
  represented as
• a very few sites that rank very high in the
  number of inbound links
• a larger number of sites with close to median
  numbers of inbound links
• a great number of sites with very few inbound
  links.

14
Network Theory and other stuff
Power Law Distribution
Logarithmic scale of distribution of inbound
links to websites categorised as Business in
the dot-com zone
Linear representation of distribution of inbound
links to websites categorized as Business in
the dot-com zone.

• Power laws as related to websites may be verbally
  represented as
• a very few sites that rank very high in the
  number of inbound links
• a larger number of sites with close to median
  numbers of inbound links
• a great number of sites with very few inbound
  links.

15
Network Theory and other stuff
Power Law Distribution
16
Network Theory and other stuff
Power Law Distribution
17
Network Theory and other stuff
Power Law Distribution
So the question is WHY?... Why is there a
power law distribution? Why is it important?
18
Network Theory and other stuff

• The web is a complex network.
• Networks (esp complex networks) exhibit certain
  characteristics.
• A characteristic of a complex net is a power law
  distribution of inbound links.
• Inbound links are also important because they are
  an element of Googles Page Rank formulation.
• Power Law distributions are said to form by the
  mechanism known as preferential attachment.
• Rich get richer
• Popularity is attractive
• Site with many links attract many links
• But what does preferential attachment mean from
  a business perspective?
• Does it bestow a competitive edge? How does one
  get it?
• How important are links?

19
Network Theory and other stuff
Power Law Distribution
So the question is WHY?... Why is there a
power law distribution? Why is it important?
20
Network Theory and other stuff
Power Law Distribution
Why is it important?
One reason is that it is in the long tail of
the power law (or the Deep Web) that much of the
Webs innovation and entrepreneurship take
place. This is where the new business models are
most likely to emerge and thus this is where they
are most likely to be discovered.
21
Network Theory and other stuff
Power Law Distribution
Why is it important?
Another reason is that as more and more commerce
move to the Web, the dynamics of link
accumulation at both global and local
granularities can strongly influence
competition and diversity in business and
society. Thus it is incumbent on future business
leaders to understand these phenomena.
22
Network Theory and other stuff
Power Law Distributions
Another reason is that as more and more commerce
move to the Web, the dynamics of link
accumulation at both global and local
granularities can strongly influence
competition and diversity in business and
society. Thus it is incumbent on future business
leaders to understand these phenomena.
So why is there a power law distribution of
Websites inbound links on the Web?
The simple answer is that popularity is
attractive.
Another suggestion is that those first adaptors
had a huge advantage and accumulated more links.
We can speculate that the phenomenon of search
engine marketing might have an impact as well.
In any event, more links indicate more popularity
which equates to greater potential for revenue
generation.
23
Network Theory and other stuff
Power Law Distributions
In any event, more links indicate more popularity
which equates to greater potential for revenue
generation.



Website plus eyeballs equals revenue (potential)