The Changing Dynamic of the Internet

The Changing Dynamic of the Internet: Early and Late Adopters of the IPv6 Standard

© Anat Hovav and David Schuff June, 2003

The Changing Dynamic of the Internet: Early and Late Adopters of the IPv6 Standard1

Anat Hovav* ([email protected])

Department of Management Information Systems Fox School of Business and Management, Temple University

207E Speakman Hall 1810 North 13th Street

Philadelphia, Pennsylvania 19122-6083 phone: (215) 204-3055

fax: (215) 204-3101

David Schuff ([email protected])

209F Speakman Hall Department of Management Information Systems

Fox School of Business and Management Temple University

1810 North 13th Street Philadelphia, Pennsylvania 19122-6083

phone: (215) 204-3078 fax: (215) 204-3101

*Corresponding author

1 Funding was provided for this research project through the Junior Faculty Grant Program at Temple University in Philadelphia, Pennsylvania.




Abstract

The United States has, for a considerable time, enjoyed a position of leadership

over the Internet. The introduction and spread of IPv6, the “next generation” Internet

Protocol, could represent a significant shift in this position. While Internet Service

Providers (ISPs) around the world are adopting IPv6, the United States has been less

aggressive. This paper, through an investigation of early and late adopters of IPv6, seeks

to understand the factors that influence the time of adoption decision. Interview data was

collected from eight ISPs in six countries. That data is applied to a set of proposed factors

derived from Roger’s characteristics of early and late individual adopters, and adapted to

organizational context. We found that those characteristics were useful in explaining time

of adoption of IPv6 for five of the eight cases we studied. Closer examination of the

remaining three cases revealed additional factors. Specifically, relative advantage,

uncertainty and risk, crisis, and power relationships also influence an organization’s time

of adoption.

Keywords:

Internet Standards adoption, Standards, IPv6, Case Study, Internet Service Providers




INTRODUCTION

A recent newspaper article (Kim 2004) cautioned that the United States is in

danger of losing its position as a technology leader. Others have voiced similar

sentiments (Butler 1992; Wrafter 2000). For example, the U.S. has lost market share to

Japan in high-tech manufacturing, office and computing machinery, and in radio and

television communications technologies (Butler 1992). Firms in the United States are no

longer the leaders in the development and adoption of mobile phones, wireless

communications, and m-commerce (Wrafter 2000). Organizations in European countries

(such as Sweden) and the Far East (such as Japan) have assumed leadership roles in these

areas.

The possibility of a similar loss in dominance over the Internet exists, due in part

to stagnation. The basic protocols used for communication over the Internet were

developed by scientists in the United States and the U.S. Department of Defense over 40

years ago. Their adoption as a global standard was in part due to the widespread adoption

of local area networks and personal computers, the use of TCP/IP1 with these platforms,

and the incorporation of TCP/IP into the UC Berkeley Unix Operating System (Leiner et

al. 1997). These protocols have remained basically unchanged (with the exception of “ad-

hoc” solutions to provide additional functionality) since their initial implementation. As a

result, the underlying technology of the Internet has not changed significantly since the

1970s.

1 Transport Control Protocol/Internet Protocol. This is the basic communication protocol used for the Internet.



The most fundamental change to the Internet proposed to date is the introduction

of a new network level protocol called Internet Protocol version 6 (IPv6). IPv6 (also

known as IPng, for “next generation”) offers a number of advantages over the current

standard, IPv4. These advantages include increased address space, mobility, auto-

configuration, multicasting, and quality of service capabilities (for additional details

regarding the key technical aspects of IPv6, refer to Appendix A).

IPv6 is being adopted extensively by ISPs in Japan and China, with other Asian

countries such as Singapore following closely behind. The European Union Commission

(EC) has mandated a timeline for the implementation of IPv6, leading to a slow but

consistent adoption of the new standard by ISPs in Western Europe. In contrast, the

adoption of IPv6 in the United States is minimal. Evidence of this disparity can be seen in

the shift in address allocation between IPv4 and IPv6 (Figure 1) – the United States

currently has 66% of the IPv4 addresses, but only 9% of the IPv6 addresses.

Allocation of IPv4 and IPv6 Addresses: Top 5 Countries

IPv4 Allocation IPv6 Allocation

United States66%

Japan6%

Canada3%

Great Britain3%

Germany3%

Rest of World19%

United States9%

Japan9%

Germany5%

Netherlands4%

Great Britian2%

Rest of World71%

Figure 1: Allocation of IPv4 and IPv6 Addresses (IPv6style.com 2004; Palet 2003)



The goal of this paper is to investigate the factors that prompt certain

organizations to be early adopters of Internet standards while others remain passive. In

the case of IPv6, we are interested in the conditions that impact time of adoption

decisions by Internet Service Providers (ISPs) operating in various regions of the world.

These ISPs serve as a logical unit of analysis because they are the population who are

making the adoption decision for the IPv6 standard. To this end, our study will focus on

the following research question: What are the factors that drive some ISPs to adopt IPv6

early, while others adopt later? Further, we attempt to provide some insight into the

apparent “late adopter” behavior of many ISPs in the United States regarding the

adoption of IPv6.

To address this question, we draw upon Roger’s (1995) characteristics of early

and late adopters. Although these characteristics focus on the individual adopter’s

personality and traits, they provide a useful framework for considering organizational

time of adoption. We establish organization-level characteristics by mapping the

individual characteristics to the literature on organizational adoption. We use the

mapping to create propositions regarding the major influences on time of adoption for

organizations. These propositions are analyzed against interviews conducted with eight

Internet Service Providers in six countries regarding their adoption (or non-adoption) of

IPv6. By examining these cases, we can learn how those individual characteristics can

influence the organizational time of adoption decision. Indeed, we find that several

factors, including sponsorship and regulation that reduce uncertainty and risk, disparities

in resource allocation leading to crisis, and the existence of “killer applications” that



provide a clear advantage to adopting the new technology, have a major impact on the

time of adoption of Internet standards.

In the next section we introduce the literature regarding characteristics of early

and late adopters as well as the relevant organizational adoption literature. In that section

we map these individual characteristics of early and late adopters to organizations.

Section three describes our methodology and the eight cases studied. Section four details

our analysis of the cases. This is followed by a discussion of the findings and future

research directions. We conclude with potential implications for industry and policy-

setting organizations.

LITERATURE REVIEW

Roger’s Characteristics of Early and Late Adopters

Most current research on the adoption of innovations in an organizational context

describes the innovation and its characteristics (e.g., Rogers 1962, 1983; Eveland and

Tornatzky 1990; Van de Ven 1991: Fichman and Kemerer 1993). For example, Rogers

(1962, 1983) proposed five fundamental characteristics of the innovation: (1) relative

advantage, (2) compatibility, (3) complexity, (5) trialability, and (5) observability.

Environmental characteristics (e.g., Farrell and Saloner 1985; Katz and Shapiro 1986;

Farrell and Saloner 1987; Fichman and Kemerer 1993; Arthur 1996) have also been

found to influence the adoption decision. For example, Katz and Shapiro (1986) and Van

de Ven (1993) discussed the positive influence of sponsorship on the adoption decision.

Generally, the outcome is considered to be dichotomous – either the organization adopts

the innovation, or it does not.



Another interesting question is the timing of the adoption decision. Rogers’

Adoption/Innovation Curve (1995) places potential individual adopters into five

categories (see Figure 2) on a continuum of time of adoption: (1) innovators, (2) early

adopters, (3) early majority, (4) late majority, and (5) laggards.

Innovators Early adopters

Early majority

Late majority Laggards

Adoption/Innovation Curve

Time

Prop

ortio

n of

ado

pter

s

Figure 2: Adoption Innovation Curve (Adapted from Rogers, 1995)

Those who adopt early (innovators, early adopters, and the early majority) are

characterized as being more “venturesome,” (Rogers 1995, p. 263) having access to

capital, can assimilate technical information, and are less risk averse. They can act as

opinion leaders, disseminating information regarding the innovation to those who have

not yet adopted. The later adopters (late majority and laggards) are more “skeptical and

cautious,” (Rogers 1995, p. 265) waiting for the innovation to become pervasive in order

to take advantage of the network externalities. They are more risk averse and less able to

financially withstand a failure due to the adoption of something new.

Organizational Characteristics of Early and Late Adopters

From these descriptions, Rogers (1995) proposed 25 characteristics that

differentiate early adopters from late adopters. Many of these also serve as plausible



differentiators of organizational adopters. We use these characteristics to form

propositions that link these “differentiators” to early and late organizational adopters by

mapping those characteristics to the literature regarding organizational adoption (see

Table 1). For example, Rogers states that individuals that are more literate are more likely

to be early adopters. This can be extended to organizations. It has been argued that

organizations with higher computer literacy (Goncalves et al. 1999) and higher levels of

technology knowledge (Iacovou et al. 1995; Chwelos et al. 2001) are more likely to be

early adopters of that new technology. Wozniak (1993) found that organizations whose

leadership is more highly educated tend to adopt innovations earlier. This leads to the

creation of an organization-level characteristic titled organizational literacy and

technical knowledge.

Rogers suggests that early individual adopters are better able to cope with

uncertainty and risk than later adopters. Similarly, Harrison et al. (1997) found that

companies that perceive more control over the adoption process (the ability to overcome

obstacles) are more likely to adopt. Lower perceptions of barriers to adopt and lower

levels of uncertainty lead to adoption by organizations (Chau and Tam 1997) while

resistance to change reduces the chance of adoption (Arvanitis and Hollenstein 2001).

Hoppe (2002) found that organizations that are potential adopters, in conditions of

uncertainty, would wait for more information. Therefore, uncertainty about the

technology leads to a lower likelihood of adoption (Luque 2002). On the other hand,

Harrison et al. (1997) found that organizations that are early adopters have a more

favorable attitude towards the new technology to be adopted, and therefore have a lower

resistance to change. Consistent with this, Chau and Tam (1997) found that companies



that are content with the technology they currently have will be less likely to adopt new

technologies.

Earlier adopters have a higher degree of communication with other organizations.

Those organizations are more highly interconnected through interpersonal networks in

their social system than later adopters. Similarly, organizations where managers talk to

managers in other companies (Huff and Munro 1985), are involved in active information

gathering (Wozniak 1993), and actively seek information from vendors (Huff and Munro

1985) are more likely to adopt new technologies.

Access to information is an important organizational determinant of adoption.

Firms that are embedded in a knowledge network that facilitates access to information

(Arvanitis and Hollenstein 2001), and firms with greater ability to search for information

are also more likely to adopt early (Hoppe 2002). The managers of organizations that are

early adopters have to engage in active knowledge acquisition (Wozniak 1993; Chau and

Tam 1997), read the general media (Huff and Munro 1985; Wozniak 1993), set a trial

phase (Huff and Munro 1985), and create test-beds when information of a new standard

is not available.

Because organizations seek information about a new innovation in order to reduce

the risk of adoption, an organization’s role among its peers can have an influence on time

of adoption. Harrison et al (1997) found that organizations that are “expected” to adopt

by their stakeholders are more likely to adopt. Firms often engage in a “waiting contest,”

refraining from adopting a new innovation until the information learned from the leaders’

experience is available to them (Hoppe 2002). These early adopters become “opinion

leaders” (Wozniak 1993).



Organizational age is interesting because the evidence of its relationship with

time of adoption is different from the effect of age for individuals. Rogers contends that

age has no effect on time of adoption – an older individual is just as likely to be an early

adopter as a younger individual. However, Luque (2002) found that older firms are less

likely to adopt an innovation because they have less opportunity to choose the latest

innovations (because they are more likely to possess an extensive existing infrastructure).

Rogers lists several characteristics of individual adopters that relate to the effect

of an individual’s wealth on time of adoption. Similarly, there is a relationship between

organizational size and slack financial resources and organizational adoption. The

availability of funding has been shown to have a positive effect on organizational

adoption (Iacouvu et al. 1995; Chwelos et al. 2001). Further, the relationship between

organizational resources and its size has been investigated in the literature. Small firms

often have trouble adopting innovations because they do not have the funding necessary

to invest in the new innovation (Arvanitis and Hollenstein 2001; Hoppe 2002; Luque

2002).

Table 1 lists the complete mapping of the individual characteristics to

organizational factors. Some of the individual characteristics cannot be easily mapped to

an organizational factor. This is not to say that these characteristics are not important. It

only indicates that there are no studies that found the factor had an influence on

organizational adoption decisions. Those characteristics with no clear mapping are

marked on Table 1 with the label “No mapping.”



Table 1: The mapping of individual characteristics to Organizational Factors

Individual (from Rogers, 1995)

Organizational Factor Sources from Organizational

Adoption Literature Socioeconomic Status Earlier adopters are more likely to be literate than are later adopters Earlier adopters have more years of formal education than later adopters

Organizational Literacy and Technical Knowledge

Wozniak 1993; Goncalves et al. 1999; Iacovou et al. 1995; Chwelos et al. 2001

Earlier adopters have higher social status than later adopters Earlier adopters have a greater degree of upward social mobility than later adopters. Earlier adopters have larger units (part of larger organizations) than later adopters

Organizational Size and Slack Financial Resources

Iacovou et al. 1995; Arvanitis and Hollenstein 2001; Chwelos et al. 2001; Hoppe 2002; Luque 2002

Earlier adopters are not different from later adopters in age

Organizational Age Luque 2002

Personality Variables Earlier adopters have greater empathy than later adopters

No mapping

Earlier adopters may be less dogmatic than later adopters

No mapping

Earlier adopters have a greater ability to deal with abstractions than do later adopters

No mapping

Earlier adopters have a more favorable attitude towards change than later adopters

Attitude Towards New Technology

Harrison et al. 1997; Chau and Tam 1997

Earlier adopters have greater rationality than later adopters

No mapping

Earlier adopters have greater intelligence than later adopters.

No mapping

Earlier adopters are better able to cope with uncertainty and risk than later adopters

Organizational Attitude Towards Risk

Chau and Tam 1997; Harrison et al. 1997; Arvanitis and Hollenstein 2001; Hoppe 2002; Luque 2002

Earlier adopters have a more favorable attitude towards science than later adopters

No mapping



Earlier adopters have higher aspirations than later adopters

Role of Organization among Peers

Harrison et al. 1997

Earlier adopters are less fatalistic than later adopters.

No mapping

Communication Behavior Earlier adopters have more social participation than later adopters Earlier adopters have a higher degree of opinion leadership

Role of Organization among Peers

Wozniak 1993; Harrison et al. 1997; Hoppe 2002

Earlier adopters are more highly interconnected through interpersonal networks in their social system than later adopters Earlier adopters have more agent contact Earlier adopters seek information about innovations more actively

Degree of Communication with Other Organizations (Competitors and Suppliers)

Huff and Munro 1985; Wozniak 1993

Earlier adopters are more cosmopolite than later adopters

No mapping

Earlier adopters have more change agent contact than later adopters.

No mapping

Earlier adopters have more access to mass media communication channels Earlier adopters have greater exposure to interpersonal communication channels Earlier adopters have greater knowledge of innovations

Access to Information Huff and Munro 1985; Wozniak 1993; Chau and Tam 1997; Arvanitis and Hollenstein 2001; Hoppe 2002

Understanding “Time of Adoption” of IPv6

Much of the current research on adoption focuses on the adoption of new

technology. There has been comparatively little research on the adoption of standards

(Lyytinen et al. 1998). In this section we use the broad categorizations created in the

previous section (and described in Table 1) to create propositions for the major influences

on early and late adoption of the IPv6 standard. Once defined, the propositions will be

verified against interview data collected from eight ISPs. Table 2 summarizes the factors



and their potential relationship to the time of adoption of IPv6. Specific information

about the characteristics of IPv6 is provided in the next section.

Table 2: Impact of Organizational Factors to Time of Adoption of IPv6 Organizational Factor Impact on the time of adoption of IPv6

1 Organizational Literacy and Technical Knowledge

ISPs in general are “technically-savvy organizations.” However some have been more exposed to IPv6 than others. Extensive knowledge of IPv6 can lead to an early adoption.

2 Organizational Size and Slack Financial Resources

Early adopters require slack financial resources to reduce the level of risk involved in being an early adopter.

3 Organizational Age Older ISPs are likely to have a larger allocation of IPv4 addresses. Also, older ISPs also have more IPv4 based infrastructure that might be incompatible with IPv6 and are thus less likely to be early adopters.

4 Attitude towards New Technology

ISPs that consider themselves innovators are more likely to experiment with IPv6.

5 Organizational Attitude Towards Risk

Organizations that are more risk averse are more like to be late adopters.

6 Role of Organization among Peers

Organizations that are considered innovators (or standard setting organizations) are more likely to be early adopters and act as opinion leaders.

7 Degree of Communication with Other Organizations (Competitors and Suppliers)

Given the distributed nature of the Internet, communication among ISPs is a fundamental requirement. ISPs need to actively look for information about IPv6.

8 Access to Information ISP’s with sufficient resources can create test environments and create knowledge about the new standard. Other ISPs rely on independent consortia for information.

Based on the above organizational characteristics of early and late adopters and

the literature on organizational adoption, we constructed six propositions regarding the

influence of organizational characteristics on the time of adoption of IPv6. Because there

is no clear theoretical direction regarding age (Characteristic 3), we have no specific

proposition for that factor. Recall that although Luque (2002) contends that older

organizations are likely to adopt later, Rogers does not see age as an influence on



individual adoption. Therefore, it is unclear what impact organizational age will have on

the adoption of IPv6. However, we include organizational age in our analysis of the cases

because of its potential relevance.

Proposition 1: Organizations with higher levels of technological savvy are more

likely to be early adopters of IPv6 (Characteristic 1).

Proposition 2: Organizations, regardless of their size, are likely to be early

adopters of IPv6 if they have enough slack funding available (Characteristic 2).

Proposition 3: Organizations that are networked and have access to information

about IPv6 are more likely to be early adopters (Characteristics 7 and 8).

Proposition 4: Organizations that have the ability to create test environments will

be early adopters and will later leverage their knowledge and become opinion leaders

(Characteristics 6 and 8).

Proposition 5: Organizations that are more risk prone are likely to be early

adopters (Characteristic 5)

Proposition 6: Organizations that are considered innovators are less resistant to

change are likely to be early adopters (Characteristics 4 and 6).

METHODOLOGY

In this study we are examining how the characteristics of individual ISPs

influence their decision to adopt IPv6. Yin (1994) suggests that exploratory studies that

try to answer questions as to “how” something is done should use case methodology.

Eisenhardt (1989) suggests that case studies may be used when little is known about a

phenomenon, or if in the early stages of research on a topic. Although adoption research

is not new, examining standards adoption is a relatively unexplored research area



(Lyytinen et al. 1998). This is especially true in the context of the Internet, and therefore

warrants the type of rich analysis this methodology can provide. For this study, we have

selected a multiple case design with a single unit of analysis for each case (also called

“type 3” case study methodology (Yin 1994)). This design can provide more compelling

evidence by supplying multiple data points by which to test theory.

Eight Internet Service Providers from six countries were used in our study. The

cases varied in size and age, serving to reduce these characteristics’ potential as sources

of bias. The subjects represent distinct regions of the world such as North America,

Western Europe, the Middle East, and Asia. We chose countries that varied in their level

of economic development, varied existing Internet infrastructure, and varied access to

communities involved in IPv6.

Within each ISP, one or more senior technical managers were selected as

interview subjects. These managers are directly responsible for infrastructure

implementation decisions within their respective organizations and therefore reasonably

represent both a managerial and technical perspective view regarding the adoption of

IPv6. In cases where more than one manager was interviewed, they were interviewed as a

group. Upon agreement to participate in the study, either face-to-face or telephone

interviews were conducted with the managers. The interviews followed a scripted set of

open-ended questions. Follow-up questions were asked when clarifications were needed.

The set of questions were developed from the list of factors derived from the literature.

The questions were phrased in such a way as to be “neutral” so that the interviewee

would not be led to answer in a particular way. For example, the question that addresses

related technologies was “What is the current availability of applications that take



advantage of IPv6?” and not “Do you agree that there are a lot of available applications

that take advantage of IPv6?”

In each case there were at least two interviewers present. One interviewer asked

the questions and recorded the responses. The second interviewer also wrote down the

interviewee’s responses to ensure that the responses were being recorded correctly. After

the interview was complete, each interview was summarized. The summaries were

compared for consistency and accuracy. Inconsistencies were resolved by follow up e-

mails or phone conversations with each interviewee. The final summaries were sent to

each subject for their review and comments. If necessary, further phone calls or e-mails

were used to clarify answers.

Table 3 lists the ISPs studied, their geographical location and their actual adoption

position.

Table 3: Internet Service Providers Used in this Study Subject Location Actual Adoption Position

1 CA North America Early 2 BI Middle East Late 3 GL Middle East Late 4 BG North America Early 5 CL Europe Early 6 GN Middle East Late 7 NX Europe Early 8 ST Asia Early

It should be noted the categories “early” and “late” as described by Rogers (1995)

are two ends of a continuum. An ISP may be considered neither an early nor a late

adopter and could fall within the “early majority” or “late majority” categories of Rogers’

adoption curve (refer to Figure 2). This could occur if the organization has some

characteristics of an early adopter and some characteristics of a late adopter. For the



purposes of this study, we have three categories of adopters – “early adopter,” “majority

adopter,” and “late adopter.”

THE CASE OF IPV6

In order to provide a richer context for analyzing the cases, the following section

describes the unique characteristics of IPv6 and why it is a particularly interesting

standard to study. We begin by explaining how we collected the information about the

state of adoption of IPv6. We continue by detailing topics such as the nature of IPv6 as an

infrastructure standard, lack of central governance sponsorship or champion, and the IPv4

address allocation issue.

To better understand these unique issues, we studied the web sites of the regional

IP address allocation agencies such as Réseaux IP Européens (RIPE), the Asia Pacific

Network Information Centre (APNIC), and the American Registry for Internet Numbers

(ARIN)2. These agencies are responsible for the allocation of Internet resources around

the world. In addition, we interviewed several individuals that are involved with the

deployment of IPv6 in various capacities. We conducted interviews with the following:

1. A senior networking engineer (from a leading networking company) that was part of

the initial design of IPv6.

2. The president of the IPv6 forum and the chair of the EU commission IPv6 Task

Force.

3. The main designer of Euro6IX.

4. Marketing Director of the North America IPv6 Task Force.

5. A member of Japan’s IPv6 Council

2 RIPE’s web site is http://www.ripe.net, APNIC’s web site is http://www.apnic.org, and ARIN’s site is available at http://www.arin.net.



6. The chair of the RIPE IPv6 work group of 6BONE registry.

IPv6 is an infrastructure technology

According to Goncalves (1999), there is a difference between the adoption of

infrastructure technologies, intermediate technologies and advanced technologies.

Infrastructure technologies (also called “architecture technologies”) are the most removed

from the user. Their value-added is difficult to clearly convey to the users (Gawer and

Cusumano 2002). IPv6 is an example of an infrastructure technology – it underlies other

technologies that take advantage of its features. However, on its own its value is difficult

to convey to the user. The user often does not know whether their ISP is using IPv4 or

IPv6. Instead, users will look for applications that require the features afforded by the

infrastructure technology.

Vendors providing the value-added applications depend on the market structure of

a given industry. In general, successfully driving the adoption of an infrastructure

technology requires either vertically integrated companies to create complimentary

technologies (e.g., IBM’s mainframe architecture) or third party vendors (e.g., Microsoft

and Intel) to collaborate and develop “killer applications.” IPv6 is not owned by anyone,

it is not patented, and there is no licensing involved. Therefore, in order to drive its

adoption, ISPs or other technology companies (e.g., hardware, software, or consumer

electronics vendors) must develop Internet-based applications that will (1) take advantage

of specific capabilities of IPv6, and therefore (2) cannot be implemented using IPv4.

For example, in the case of the “Wintel” standard, new versions of Windows (the

application) take advantage of the advances in Intel processors (the infrastructure). In the

case of the mainframe, IBM introduced applications that took advantage of their own new



hardware. However, in both cases the technology was championed by the vendor (e.g.,

Intel and IBM). In contrast, IPv6 was proposed and developed by an independent,

voluntary group (IETF) and is not championed by any single company. As of 2004, IPv6

is being promoted by the IPv6 forum and regional and national task groups. In the words

of the Marketing Director of the North American IPv6 task force, “there is no one entity

that is out to make money of the introduction and the adoption of IPv6.” Instead, profit

can only be made from the introduction of complementary technologies that will take

advantage of the new standard.

Relatively little information exists regarding IPv6

One of the issues facing early adopters is the risk associated with the adoption of

a new standard. The risk is a result of uncertainty as to the future of that standard. An

adopting organization can reduce its risk if they can leverage their knowledge about the

new standard to encourage its adoption. In the case of IPv6, the lack of a single champion

and a lack of a mandate from a central governing body increase the uncertainty and risk

associated with its adoption.

In many cases the developing vendor supplies early adopters with information

regarding the features of the technology, technical specifications, and training (Gawer

and Cusumano 2002). In the case of IPv6, adopters are forced to rely more heavily on

their own test beds (which require significant investment), internal training, trade

publications, or support from consortia and governments.

Because IPv6 was developed by the Internet Engineering Task Force (IETF)3 (a

voluntary consortium with limited funds), there is no significant marketing effort

encouraging IPv6 adoption. This is in sharp contrast to Intel, which used the “Intel 3 More information about the IETF is available through its website at http://www.ietf.org.



Inside” campaign to promote its brand name. IPv6 is already embedded in Cisco routers,

Microsoft’s Windows XP operating system, and some Nokia phones. However, most

companies do not publicize this support for IPv64. In addition, the amount of information

available in trade publications is limited. For example, a survey of four top trade

magazines (see Appendix B for the search methodology used) revealed that the number

of articles regarding IPv6 between 1998 and 2004 totaled 97, compared to 743 articles

about Windows XP and approximately 1,500 articles about XML (both championed by

Microsoft). Although the number of articles increased from a few (between five and nine

per year) in the years 2000 through 2003 to about 40 in 2004, it is still a fraction of the

number of articles published regarding the other standards. This lack of information can

lead to uncertainty regarding the value of IPv6 and the complexity of its implementation.

In some regions there are major efforts to create test beds and distribute

information on IPv6. Such efforts are the 6BONE and the Euro6IX. These efforts are

mostly concentrated in Europe. In addition, the IPv6 forum conducts information sessions

in various regions of the world. According to the President of the IPv6 forum, the

attendance of these formal information sessions in the Far East is overwhelming. In the

United States, however, attendance is very low. Overall, he stated, there is very little

interest in seeking out information about IPv6 in North America.

IPv6’s lack of a champion and sponsorship

As mentioned above, there is no private sponsorship for IPv6. None of the major

Information Communication Technology companies have adopted IPv6 as its platform or

is advocating its adoption as the next Internet Protocol. Also, government sponsorship for

IPv6 is generally limited, varying by region. In Asia, there are various levels of 4 Although an IPv6 ready logo program was initiated in 2003, it is limited in scope and very low in funding.



sponsorship depending on the country. Japan, for example, provides tax incentives and

has invested over eight million yen in the promotion of IPv6. In Pakistan, the government

provides some training. The European Union (EU) has mandated the implementation of

IPv6 as a long-term goal to increase the competitive position of the European community.

The EU also provided over 180 million Euros for research and development. However,

the mandate is not accompanied by financial incentives for the commercial sector (i.e.,

ISPs). In North America, however, there are no financial or regulatory incentives.

IPv6 solves the IPv4 “address crisis”

The concept of “crisis” is not new to Information and Communication

Technologies. Microsoft consistently forces upgrades by discontinuing support and

maintenance of older products. This practice also forces complementary technology

upgrades (i.e., faster processors to run the newest version of Windows). In the case of

IPv6, if all major networking component providers announce the discontinued support of

IPv4, they could create this type of crisis, forcing ISPs to upgrade their networking

equipment to be IPv6-compatible. However, since there is no single sponsor of the

technology, vendor-induced “forced crisis” does not occur.

However, there is another form of crisis created by the limitations of the current

IPv4 standard. Specifically, IPv6 solves an intractable problem of the current IPv4

standard – the allocation and availability of network addresses. The number of IP

addresses currently available under IPv4 is limited and fixed. The result is that IPv4

addresses are a scarce resource with limited growth capabilities. Further, the class

structure of current IP addresses created an uneven distribution of addresses. For

example, a class A address, when allocated to an organization, can hold approximately



two million addresses. An organization that receives a class A address has control over

these addresses, whether they use them or not.

IPv4 addresses were allocated on a “first come-first serve” basis, and not on the

basis of need. This also creates a highly uneven distribution of addresses. The United

States and Canada own over 70% of all IPv4 addresses, Europe owns a little over 20%

and the rest of the world owns less than 10%. Xerox (which was allocated a class A

address in 1991) currently owns more addresses than the entire Republic of China (which

owns about 31,000 addresses). The result is a major concentration of a scarce and limited

resource over a relatively small population.

The impact of organizational age

As mentioned above, the effect of organizational age on time of adoption is

unclear. Older organizations traditionally have more investments in existing

infrastructure and thus are less likely to be early adopters. At the same time, older

organizations usually have more funding, skills and experience and are thus better

positioned to absorb the risk involved in early adoption. In the case of IPv6, age has two

additional important attributes. Newer ISPs are more likely to have equipment that is

already IPv6 compatible, making the upgrade less costly and less complex. Likewise,

since IPv4 addresses were allocated based on a “first come, first served” scheme, newer

ISPs are more likely to have a resource deficiency. Age is not independent from the other

characteristics discussed in this section – instead it is tied closely to the cost to upgrade

and available funds, and crisis. Much like other ISPs, newer ISPs are more likely to adopt

IPv6 early if they have sufficient funds or if they are facing an address crisis.



In the next section we analyze the eight case studies in light of the propositions

we introduced previously and in the context of the unique characteristics of IPv6.

ANALYSIS OF CASE STUDIES

Subject 1 – CA

Although not a traditional private Internet Service provider, CA provides Internet

connectivity for a major Canadian University and eight other organizations, including

hospitals, other universities, and research institutions. CA is a highly “technical savvy”

organization. In order to gain this knowledge, CA had to acquire knowledge about IPv6

early. This knowledge was acquired through participation in a community network of

adopters such as 6NET and CANARIE’s CA*net, a national optical network for research

and development. They also gain knowledge through the creation of test environments,

facilitated by the ample slack resources available to the organization. Although they see

stability as an important aspect of their overall strategy, their mission is one of

innovation, as they are charged with servicing a research environment. Their mission

compels them to offer IPv6, regardless of the risks associated with being an early adopter.

Based on Proposition 1 through 6, CA has the characteristics of an early adopter.

CA has taken a leadership role in offering IPv6 to their clients thus acting as an

early adopter. They explicitly stated that they were not waiting for other organizations

to implement the new standard, and they will provide IPv6 connectivity to any client who

requests it.

Subject 2 –BI

BI is one of the top five Internet Service Providers in its market (out of about 40

ISPs). BI is a subsidiary of the national telecommunications company, which historically



has been public, but is currently being privatized and therefore has control over most of

the infrastructure in its country. BI is a highly bureaucratic organization and will try to

protect as much of its investment in the existing infrastructure. As a result, BI resistance

to change is high, preferring stability to innovation. BI has little technical knowledge

regarding IPv6. They are not involved in any IPv6-related community, they get most of

their information from trade magazines, and they do not have funds to create test beds. BI

perceives the implementation of new technologies as risky, increasing the levels of

uncertainty associated with the implementation of IPv6. In addition, BI does not have the

slack resources required to fund highly uncertain projects. Based on Propositions 1

through 6, BI has the characteristics of a late adopter.

BI has taken a late adopter position in regards to IPv6. They have no plans to

implement IPv6 in the foreseeable future. They did not study the new standard or its

benefits, nor do they know what will be the cost to upgrade.

Subject 3 - GL

GL and BI are operating in the same region, and GL has similar characteristics to

BI with the exception of its age and position in the market. This ISP is relatively new to

the market. GL has been providing international telephony and data services in its market

for the last five years, but their ISP division has been in existence for less than two years.

While BI is the incumbent, GL is a new entrant trying to gain a market share. They define

themselves as leaders in services but not in the implementation of infrastructure.

However, their mission is to become the leader in their market by providing better

services then the more established ISPs. To facilitate this, GL is less bureaucratic and has

lower resistance to change. Much like BI, GL has little technical knowledge regarding



IPv6. Although they have some awareness of the limitations of IPv4, they do not have

extensive knowledge of the features of IPv6. They are not involved in any IPv6-related

community and they get most of their information from trade magazines. GL would like

to be able to experiment with new technologies by setting up test environments but they

do not have the funds to do so. Therefore, they closely follow opinion leaders (mostly in

the United States). Based on Propositions 1 through 6, GL has the characteristics of a late

adopter.

Despite GL’s mission to become a service leader, their approach to the adoption

of IPv6 is that of a late adopter. They do not plan to implement IPv6 in the foreseeable

future. However, GL’s approach appears not to be in their best interest. They

acknowledge the existence of a major shortage in IP addresses and they describe the

difficulties in getting new addresses and the effort and time requires in managing the

addresses they have. However, they have made no attempt to study the addressing

possibilities of IPv6. In addition, GL follows the United States in their adoption decision

regardless of the disparities in the allocation of IP address that exists between ISPs in the

United States and ISPs in the Middle East.

Subject 4 – BG

BG is a technology group in a North American government agency servicing

internal clients. They currently support an extensive IPv4 infrastructure and are a part of

a highly bureaucratic organization. Therefore, their resistance to change is expected to be

high. At the same time, part of BG’s mission is to test and evaluate new technologies for

their client base.



BG is highly technically savvy organization. As of 2002, BG maintained an IPv6-

based test environment with seven sites and about 40 users with an expected growth of

four to five sites a year. These sites are one of the very first test beds created for IPv6.

BG is also involved with the North American IPv6 taskforce and therefore has access to

information about the new standard. In addition, BG’s test beds have been used to study

issues such as the cost of upgrading from IPv4 to IPv6, compatibility and transitioning,

and security. As a result, BG acts as an opinion leader. The risks associated with early

adoption are irrelevant to BG due to the extensive funding they receive as a government

agency. As a result, BG has enough slack resources to fund test beds.. Based on

Propositions 1 through 5, BG has the characteristics of an early adopter, although

Proposition 6 provides some evidence that they might be more reluctant to adopt early

(due to their bureaucratic nature and high resistance to change).

BG considers themselves an early adopter of new technologies and they place

more importance on innovation than consistency of service. They did not wait for other

ISPs to implement IPv6, but they are aware of the need for interoperability among ISPs.

What appear to drive BG to implement IPv6 are its features and the potential for new

applications or the relative advantage they can gain from an early implementation of the

new standard.

Subject 5 – CL

CL is a small, European consulting company that is providing end-to-end

solutions using IPv6. CL began as a networking and Internet consulting company for

Internet Service Providers and other companies. They began to offer connectivity

services recently, when IPv6 was first introduced in their country. CL defines themselves



as leaders in the implementation of technology. They have already implemented IPv6.

Some of their clients (university and educational institutions) are already on IPv6. They

expect 80% or their other clients to be using IPv6 within two years.

Most of CL’s equipment is compatible with the new standard and the equipment

that is not compatible can be connected to an IPv6 network. CL is a relatively small and

versatile company thus, they are accepting of change. However, they do not have large

slack funds. Nevertheless, CL believes that it is more risky to ignore IPv6 than to

implement it. In addition, they do not believe the cost of implementation to be

prohibitive. CL is highly technically savvy organization. They are heavily involved with

the IPv6 forum and Euro6IX and act as opinion leaders in the European community. CL

has the characteristics of an early adopter.

However, CL sees the implementation of IPv6 as a necessity to remain

competitive. CL therefore has implemented IPv6 extensively, acting as an early adopter.

The main reason CL has adopted the new standard is because of its features. CL is

offering home automation and mobile computing services. The auto-configuration,

mobile capabilities, and the ability to implement Universal Plug and Play allow CL to

design and offer new applications and services. A second driver is the cost of IP

addresses. CL operates in an environment with relatively few addresses. Buying

additional IPv4 addresses was very costly. CL preferred to invest those funds in the

adoption and implementation of IPv6.

In the first five cases the characteristics of the ISPs match their adoption position,

supporting Rogers’ characteristics of early and late adopters and their general



applicability to an organizational context. For example CA’s characteristics suggest that

CA will act as an early adopter and they are indeed early adopters. BI and GL, on the

other hand, have the characteristics of later adopters and they are indeed later adopters.

Yet, BI has a reason to behave as a later adopter. BI is the incumbent ISP in their country

and they are trying to protect their investment in the current infrastructure. GL’s behavior

as a late adopter might appear counter intuitive, but its lack of knowledge and resources

to set up test environments force them into following an opinion leader which operates

under a different set of constraints than they have. With the exception of their resistance

to change, BG has the characteristics of and behaves as an early adopter. In the case of

CL, their lack of slack funds doesn’t prohibit them from becoming an early adopter of

IPv6 because they do not want to miss the opportunity to leverage the capabilities

associated with IPv6 into value-added services.

In the remaining three cases, the implied adoption position does not completely

match the actual time of adoption decisions that were made.

Subject 6 – GN

GN provides mobile Internet access. It provides services to handsets, personal

digital assistants, and laptops with wireless modems using wireless technologies such as

the Wireless Application Protocol (WAP). GN’s user base is ten times larger than the

next largest mobile ISP in their market. Their strategy is to maintain their position as a

market leader by adopting innovative technologies while maintaining a stable

infrastructure. Therefore, GN’s aversion to risk is relatively high. GN defines themselves

as leaders in services but not in the implementation of infrastructure. Being a relatively



new ISP and servicing a less traditional market, GN’s resistance to change is relatively

low.

GN studied the issues surrounding the implementation of IPv6. They are aware of

the features afforded by the new standard and its capabilities. They monitor providers in

the United States, and their adoption decisions are influenced by major vendors and

global ISPs (such as Cisco, UUNET and the NAPs). However, GN is not involved in any

IPv6-related community, consortium, or task force, nor do they have the funds to create

test beds. GN does not have excessive financial resources, but they state that the cost to

upgrade from IPv4 to IPv6 is minimal since most of their equipment is IPv6 ready. They

maintain that cost is not the reason they are not implementing IPv6, but acknowledge that

the cost to maintain backwards compatibility is an issue. Based on propositions 1 through

6, GN has characteristics that make them a majority adopter.

Although GN studied the implementation of IPv6, they do not plan to implement

IPv6 in the foreseeable future, making them a late adopter. This statement, however,

needs to be qualified. GN follows the United States in terms of Internet standards. Yet,

GN services a mobile Internet market that uses European rather than North American

wireless standards. In the event that the newly proposed 3G wireless standard prevails,

GN will have to implement IPv6.

In the case of GN, much like in the case of BG, the features of the new standard,

the ability to create new applications and the competitive advantage gained from the

implementation of the standard are the main drivers for its adoption. The difference

between the two cases is that BG has the resources to test the technology and drive the



market while GN is taking a more passive role, watching the market trends and waiting

for the right time to implement.

Subject 7 – NX

NX is a small, European ISP with about 3,000 customers. It is the third oldest ISP

in its market, having been in business since 1992. The company offers both traditional

ISP services and consulting services. NX defines themselves as leaders in the

implementation of technology. As a service provider they believe that advance

knowledge is important and can provide them with a competitive advantage. Thus, they

are accepting of change and are not risk averse. NX is a technically savvy organization.

They are a small ISP and have limited funding, yet they are testing the implementation of

IPv6 (with servers that run both IPv4 and IPv6) internally. NX is planning a full

implementation (of native IPv6 servers) in the near future but they will need to rely on

financial aid and intend to apply to their government and 6BONE. NX’s small size and

their financial limitations combined with their technical knowledge and their low risk

aversion position them as a majority adopter.

However, NX sees IPv6 as an opportunity to offer “turnkey” services for

organizations transitioning to the new protocol. They expect clients to start requesting

IPv6 in the near future. In their plans to adopt IPv6, they act as an early adopter. The

main reason NX believes in the viability of IPv6 is because the European Union (EU)

supports the adoption of the standard. According to NX management, the EU believes

that IPv6 is essential for the future development of Europe. The EU introduced a 10-year

telecommunications mandate in which the adoption of IPv6 has a key role. This mandate



ensures the acceptance of IPv6 as a standard and makes an individual’s aversion to the

risk involved with being an early adopter irrelevant.

Subject 8 - ST

ST is an ISP located in Asia. Their customers include both private users and small

and large corporate accounts. ST has been in business for three years - during that time,

the size of their market has tripled. ST defines themselves as followers and state that

consistency of service is more important than innovation. Thus, they are risk averse and

prefer stability over radical change. ST does not have the financial resources to invest in a

major upgrade. However, they state that the cost to upgrade is not a major issue since

most of the technology they use is already compatible with IPv6. ST can be considered

somewhat technologically savvy. They have a sufficient knowledge of the features

afforded by IPv6 but they lack information regarding implementation, such as the length

of time to complete the upgrade and the amount of training that will be required

(although the training will be free, provided by their government). ST is not involved

with IPv6-related communities, consortia, or task groups (the IPv6 forum does not have a

representative in their country). Considering ST’s characteristics and Propositions 1

through 6, ST should be a late adopter.

However, ST behaves as an early adopter. They are planning to start the

installation of IPv6 within six months and complete it within a year. They intend to

implement IPv6 in the near future because they are facing a major address shortage. ST

states that getting addresses is a serious problem. They must demonstrate that they have

run out of addresses before they can request additional ones. In addition, the price of a



block of IP address has doubled in the last few years. Given their current growth, they

will need to quickly acquire additional addresses.

Unlike the other ISPs, ST is defines itself as a late adopter. Yet the lack of IP

addresses in their region forces them to adopt the new standard. This is despite lacking

the slack financial resources available to CA or BG, and without the extensive

government support available to NX.

Table 4 summarizes the attributes of the eight cases, the implied time of adoption

based on the six propositions, and their actual time of adoption based on our interview

data. It is evident from this analysis that in some cases our propositions regarding the

factors’ prediction of early and late adopters were confirmed, while others were not. In

the next section, the implications of this analysis will be discussed.

DISCUSSION

The first five cases listed in Table 4 include two late adopters and three early

adopters. In these cases, the implied time of adoption based on the ISPs characteristics

and the six propositions match their actual adoption time. This provides support for our

contention that applying Rogers’ characteristics of early and late adopters is also useful in

an organizational context.

However, the adoption decision of the remaining three cases did not match the

outcome that was predicted using Rogers’ attributes. Two of the ISPs were predicted to

be in the “middle majority” based on their characteristics. Interestingly, one of them

acted as an early adopter (NX) and the other as a late adopter (GN). The final case (ST)

was the most surprising. ST’s characteristics were that of a late adopter but ST behaved

as an early adopter of IPv6. One explanation for our findings is that there might be



additional factors that influence an organization’s time of adoption. To better understand

what these additional factors might be, we revisited the unique characteristics of IPv6

described previously and the data collected in the cases.

Table 4. Summary of the Analysis Evidence of Factor by Subject

Propositions/ Factors

CA BI GL BG CL GN NX

ST

P1 Technological savvy

Yes No No Yes Yes Yes Yes No

P2 Slack resources

Yes No No Yes No No** No No

P3 Access to information

Yes No No Yes Yes No Yes No

P4 Ability to create test environments

Yes No No Yes Yes No No*** No

P5 Low aversion to risk

Yes No No Yes Yes No No No

P6 Innovativeness/ Low resistance to change

Yes No Yes No Yes Yes Yes No

Implied time of

ISPs adoption of IPv6

Early Late Late Early Early Majority Majority Late

Actual time of ISPs adoption of IPv6

Early Late Late Early Early* Late Early Early

Propositions supported

All All 1, 2, 3, 4,

5

1, 2, 3, 4,

5

1, 3, 4, 5,

6

2, 3, 4, 5

1, 3, 6 None

*CL is the only company we talked to that is providing IPv6 services commercially **Although GN has little excess funding, the cost to upgrade is not an issue ***NX has technically done some testing, but it is extremely limited.

The existence of a killer application

One of the unique features of IPv6 is that it is an infrastructure technology,

removed from the user, thereby requiring a “killer application” to drive adoption.

Therefore, we expect that an early adopter will be:



1. An ISP that is vertically integrated and provides end-to-end Internet solutions (such

as CL). In this case the ISP is likely to benefit from the IPv6-related applications it

offers.

2. Where there are “killer applications” being developed. For example, European

telecommunications companies have recently introduced 3G phones, which rely

heavily on IPv6. GN stated that if the 3G wireless standard prevails, they will adopt

IPv6 (this is because GN’s wireless services depend on the European wireless

standards). The constraint of a “killer application” can explain GN’s behavior. They

realize that if Europe moves to 3G, they will have to upgrade to IPv6 quickly.

Evidence of this is seen in their high level of technical knowledge and their existing

infrastructure, which is already IPv6-compatible. However, they are declining the role

of a “first mover,” since they do not have the funds necessary to create test beds and

are highly risk averse.

In summary, the relative advantage that can be gained from the introduction of a

new technology or standard can encourage speedier adoption by organizations that might

otherwise adopt late.

The impact of uncertainty and risk

The risk involved in the adoption of IPv6 is exacerbated by the lack of central

governance or a private champion. This risk is associated with the levels of uncertainty

surrounding the new standard. Under these circumstances, first movers will adopt the

new standard if:

1. They have ample slack financial resources and the investment in the new technology

is so minimal compare to the potential gains that it is worth the risk (i.e., CA, BG).



2. The levels of uncertainty can be offset financially through government sponsorship.

This support can reduce the economic risk involved in early adoption.

3. Government regulations can also reduce the level of risk and uncertainty about the

future of a new standard. If a government mandates the implementation of a new

standard, that standard will become dominant (at least in that country or region),

ensuring some market of related products. This is consistent with the behavior of NX.

Despite NX’s implied position as a “majority” adopter, they are adopting IPv6 early.

NX believes that the European Union mandate will force the widespread adoption of

IPv6 – in this environment, by adopting early they are positioning themselves to be

opinion leaders.

Another aspect of uncertainty is the pervasiveness of information about the new

standard. IPv6’s lack of a champion, combined with the limited marketing capabilities of

the IETF and the IPv6 forum, has resulted in a limited amount of information available to

adopters. ISPs that are involved in private consortia such as 6BONE, Euro6IX and

CANARIE (for example, CA, BG, and NX) have greater access to information. Others

(for example, BI and GL) have less access leading to uncertainty about the technology.

In summary, we must also consider certain environmental factors when

considering the time of adoption in an organizational context. Organizations that are in

the middle majority might become early adopters if they have financial or regulatory

support from their governments (e.g., NX), while organizations that do not have such

support will tend to remain in the middle or become late adopters (e.g., GN).



The impact of crisis

A third factor that may affect organizational time of adoption is crisis. As

previously discussed, crisis could be externally induced (such as a vendor discontinuing

product support or a government mandate) or inherent in the technology (such as the

limited number of possible IPv4 addresses). ISPs might adopt a new standard, even if

there is uncertainty as to its success, because they face a crisis of sufficient severity.

The existence of a crisis explains the behavior of ST. ST has the characteristics of

a late adopter. As a relatively small ISP with limited financial resources, it is risk averse

and does not consider itself a leader. However, it is acting as an early adopter. ST stated

that the main reason for their adoption of IPv6 is a lack of IPv4 addresses and the related

sharp increase in the cost of those addresses. In 2000, the cost of a “class C” address was

between $1,050 and $1,275 per year. In 2002, the cost was between $1,900 and $2,300 a

year.

Interestingly, ISPs in the Middle East (especially GL and GN), which rely on the

European address pool do not perceive a crisis at this point. They acknowledge that

getting addresses is becoming more difficult (and somewhat more costly) but they do not

consider their lack of addresses to be at a crisis level. However, GL and GN are in a

country that, in the short term, has a sufficient supply of IPv4 addresses (one IPv4

address for every two people). This is in strong contrast to ST, where addresses are in

short supply (it has only has one IPv4 address for every 730 people). Thus, it appears

that difficulty in obtaining new IP addresses does not always lead to perception of a

crisis.



Although forced crisis has been a common issue in information technology, there

is no forced crisis related to IPv6. The EU mandate is a long-term transitional plan, and

therefore does not introduce a crisis. In addition, none of the major networking

component providers have discontinued (or announced plans to discontinue) IPv4

support.

The impact of resource concentration and power

Organizations in the United States have control over a significant majority of

available IPv4 addresses. IPv4 addresses must be unique and there are a finite number of

them, making them a scarce resource. As the need for IP addresses escalates, the price of

each address will increase, in turn increasing their value to the companies that currently

own them. From a resource allocation perspective, the United States has little incentive to

promote the adoption of IPv6. This introduces a fourth factor, power. As stated in Hart

and Saunders (1997), power relationships can impact organizational adoption decisions.

In this case, ISPs in Europe and the Far East might drive the adoption of IPv6 in an

attempt to equalize control over the Internet. To quote the European Union commission

report: “The risk of IPv4 addresses becoming increasingly scarce by 2005, coupled with

the uneven distribution of address space between North America and the rest of the

world, is sufficiently serious for action to be taken now and swiftly...” (Communications

of the European Communities 2002, page 7)

Figure 3 presents a summary of the predictors of the time of adoption of IPv6. We

found that in addition to organizational characteristics, the degree of relative advantage,

uncertainty and risk, crisis, and power relationships all may influence the time of

adoption.



OrganizationalCharacteristics

(Roger’s characteristicsof individual adopters)

Time ofAdoption

(Early or late)

Prop

ortio

n of

ado

pter

s

Time

Framework for Predicting Time of Adoption for IPv6

Relative Advantage

Uncertainty andRisk

Crisis

PowerRelationships

Figure 3: Framework for Predicting Time of Adoption for IPv6

LIMITATIONS AND FUTURE RESEARCH

This study has two limitations. First, although IPv6 is an interesting case for

examination, the factors that we found relevant to its time of adoption may not apply to

other standards. For example, the limited number of IPv4 addresses is a unique problem.

Second, attempting to predict whether an ISP would be an early, majority, or late adopter

of IPv6 was only straightforward (see Table 4, page 28) when all six factors indicated a

single outcome (for example, in the case of CA, all factors indicated that the ISP would

adopt IPv6 early). Predicting the time of adoption when the factors were directionally

opposite was more difficult, and resulted in the outcome being predicted by a careful

examination of the case data (instead of, for example, a simple count to find a majority).

Because this analysis, in part, relied on the authors’ judgment, it is possible that this did

not accurately capture the relative weights of each factor, resulting in inaccurate

predictions of the time of adoption.



Future research should focus on addressing these limitations. The six factors we

derived from Roger’s characteristics of early and late adopters could be applied to other

standards (along with the additional four factors we discovered through our analysis).

Also, determining the relative weight that should be given to each of these factors would

increase our level of understanding of the relative importance of each factor. In addition,

the perspective of this study has been from the adopting organization (in this case, the

ISP). Future studies should consider this issue from varying perspectives, such as vendors

(of infrastructure technologies and advanced technologies), the development community,

the end user, and regulators and policy makers.

Another interesting lens through which to study the timing of Internet standards

adoption is to more fully investigate the role of power and control over resources. This is

especially interesting in the case of IPv6 because of the scarcity of IPv4 addresses.

However, this has implications for all Internet standards because of the widely varying

levels of sponsorship, combined with the Internet’s overall lack of central governance.

CONCLUSIONS

The purpose of this study was to more clearly understand the adoption of the

Internet-based protocol standard IPv6. Specifically, we sought to learn why some Internet

Service Providers adopt that standard earlier than others. To address the problem, we

investigated the factors that drive an ISP’s time of adoption? To that end, we adapted

Rogers’s characteristics of early and late adopters to organizations. Our study also found

that in addition to these characteristics, the following factors may impact organizational

time of adoption of a new standard: (1) the relative advantage associated with the new

standard, (2) the levels of uncertainty and risk involved with being an early adopter, (3)



the existence of crisis, and (4) the power relationships resulting from control over the

existing standard.

The fourth factor, power relationships, provides some insight into the apparent

“late adopter” behavior of many ISPs in the United States regarding IPv6. One might

argue that the majority of ISPs in the United States are not acting as early adopters from a

strategic and power position. Given the considerable control that organizations in the

United States have over the current pool of IPv4 address, it is to their advantage to

maintain the status quo.

If that trend continues, the possibility exists for the emergence of two “Internets,”

one based on IPv4 and the other based on IPv6. This is most likely to happen if strong

forces in Europe and Japan drive an absolute implementation of IPv6 (also known as

native IPv6) while ISPs in the United States remain with their IPv4-based networks. The

existence of two Internets will require the implementation and maintenance of

transitional technologies (networking components that will translate between the two

protocols) and conversion points where the two networks are connected. The need for

these transitional technologies will result in increasing cost to maintain a global Internet

and the existence of highly concentrated points of failure resulting in communications

failures between the two networks. If maintenance costs in terms of funds and

coordination become prohibitive there might be a complete breakdown of the global

information superhighway.

In addition, services that are available to customers on the IPv6 Internet may not

be available to customers on the IPv4 Internet (one such service is high-speed, always-on,

mobile Internet connectivity via a cellular phone). This disparity will increase as greater



numbers of “killer applications” for IPv6 are introduced. This increasing difference in

service levels between the two networks could eventually cause the United States to lose

its dominance over the Internet.

Therefore, we see two alternatives for organizations in the United States looking

to retain their key role in the development of the Internet: either the development of an

IPv6-based killer application or government intervention. This intervention could take

one of two forms, both of which reduce the risk involved in early adoption. The first is a

mandate (as can been seen in the case of the European Union). The second is a partial

subsidization of some aspect of adoption (as can be seen by the funding of training in the

ST case, or by the tax incentives given in Japan).

REFERENCES

Arthur, W. B.“Increasing Returns and the New World of Business,” Harvard Business Review, (74:4), 1996, pp. 100-109.

Arvanitis, S., and Hollenstein, A.H. “The Determinants of Adoption of Advanced

Manufacturing Technology – An Empirical Analysis Based on Firm-level Data for Swiss Manufacturing,” Economics of Innovation and New Technology, (19:5), 2001, pp. 377-414.

Butler, A. “Is the United States Losing its Dominance in High-Technology?” The Federal

Reserve of St. Louis Review, (74:6), 1992, pp. 19-34. Chau, P.Y.K., and Tam, K.,Y. “Factors Affecting Adoption of Open Systems: An

Exploratory Study,” MIS Quarterly, (21:1), 1997, pp. 1-24. Chwelos, P., Benbasat, I., and Dexter, A. “Research Report: Empirical Test of an EDI

Adoption Model,” Information Systems Research, (12:3), 2001, pp. 304-321. Commission of the European Communities. “The Next Generation Internet – Priorities

for Action in Migrating to the New Internet Protocol IPv6” [online], 2002. http://europa.eu.int/eur-lex/en/com/cnc/2002/com2002_0096en01.pdf [last accessed 03/30/2004]

Eisenhardt, K. “Agency Theory: An Assessment and Review,” Academy of Management

Review, (14:1), 1989, pp. 57-74.



Eveland, J. E. and Tornatzky, L. G. “The Deployment of Technology,” In: The Processes

of Technology Innovation, L.G. Tornatzky and M. Fleischer (eds.), Lexington Books, Lexington, Massachusetts, 1990, pp. 117-148.

Farrell, J. and Saloner, G. “Standardization, Compatibility and Innovation,” Rand Journal

of Economics, (16:1), 1985, pp. 70-83. Farrell, J. and Saloner, G. “Competition, Compatibility, and Standards: The Economics

of Horses, Penguins and Lemmings.” In: Product Standardization and Competitive Strategy, H. L. Gabel (ed.), Elsevier Science, Amsterdam, 1987, pp. 1-21.

Fichman, R. G. and Kemerer, C. F. “Adoption of Software Engineering Process

Innovations: The Case of Object Orientation,” Sloan Management Review, (34:2), 1993, pp. 7-22.

Gawer, A. and Cusumano M.A. Platform Leadership: How Intel, Microsoft and Cisco

Drive Industry Innovation, Harvard Business School Press, Boston, Massachusetts, 2002.

Goncalves, V., Palma-Dos-Reis, A., and Duque, J. “Portuguese Financial Corporations’

Information Technology Adoption Patterns,” Interfaces, (29:4), 1999, pp. 44-57. Hart, P. and Saunders, C. “Power and Trust: Critical Factors in the Adoption and Use of

Electronic Data Interchange,” Organization Science, (8:1), 1997, pp. 23-42. Harrison, D.A., Mykytyn, P.P., and Riemenschneider, C.K. “Executive Decision About

Adoption of Information Technology in Small Business: Theory and Empirical Tests,” Information Systems Research, (8:2), 1997, pp. 171-195.

Hoppe, H.“The Timing of New Technology Adoption: Theoretical Models and Empirical

Evidence,” The Manchester School, (70:1), 2002, pp. 56-76. Huff, S. L.M., and Munro, M. C. “Information Technology Assessment and Adoption: A

Field Study,” MIS Quarterly, (9:4), 1985, pp. 327-339. Iacovou, L.C., Benbasat, I., Dexter, A. “Electronic Data Interchange and Small

Organizations: Adoption and Impact of Technology,” MIS Quarterly, (19:4), 1995, pp. 465-485.

IPv6style.com. “Statistics: IPv4 Address Assignment Ranking” [online], 2003.

http://www.ipv6style.jp/en/statistics/ipv4assign/ [accessed March 27, 2004].



Katz, M. L. and Shapiro, C. “Technology Adoption in the Presence of Network Externalities,” Journal of Political Economy, (94:4), 1986, pp. 822-841.

Kim, J. “U.S. Could Lose Technology Dominance, Executives Say,” Washington Post,

2004, January 8, page E05.. Leiner, B., Cerf, V., Clark, D., Kahn, R., Kleinrock, L., Lynch, D., Postel, J., Roberts, L.,

and Wolff, S. “The Past and Future History of the Internet,” Communications of the ACM, (40:2), 1997, pp. 102-108.

Luque, A. “An Adoption-Value Approach to Technology Adoption in U.S.

Manufacturing: Evidence from Microdata,” Economics of Innovation and New Technology, (11:6), 2002, pp. 543-568.

Lyytinen K., Rose G., and Welke R. “The Brave New World of Development in the

Internetwork Computer Architecture (InterNCA): or How Distributed Computing Platforms will Change Systems Development,” Information Systems Journal, (8:3), 1998, pp. 241-253.

Palet, J. Presentation: European IPv6 Task Force Status, IPv6 Task Force Steering

Committee, 2003. Rogers, E.M. Diffusion of Innovations, The Free Press, New York, 1962. Rogers, E.M. Diffusion of Innovations (2nd Edition), The Free Press, New York, 1983. Rogers, E.M. Diffusion of Innovations (4th Edition), The Free Press, New York, 1995. Van de Ven, A. H. (1991). Managing the process of organizational innovation, In:

Changing and Redesigning Organizations. Huber, G.P. (ed.), Oxford University Press, New York.

Wrafter, M. “Europe Threatens U.S. Internet Dominance. Vision Consulting.” [online],

2000. http://www.markwrafter.com/groove/penned/vision/00/060300.html. [Accessed March 22, 2004]

Wozniak, D. G.“Joint Information Acquisition and New Technology Adoption: Late

Versus Early Adoption,” Review of Economics and Statistics, (75:3), 1993, pp. 438-445.

Yin, R. K.Case Study Research Design and Methods. Sage Publications - Applied Social

Research Methods Series, Thousand Oaks, California, 1994.



Appendix A: Comparison of IPv6 to IPv4

Category Advantage of IPv6 Why it is Important Addressing The address space in IPv6 is much

larger than IPv4 (16 bytes instead of 4 bytes). This means that IPv6 allows for 3.4 x 1038 addresses, compared with 4.2 x 109 possible addresses with IPv4.

The number of unique IPv4 addresses is dwindling rapidly. This is mostly a problem in undeveloped countries.5 It is also anticipated to become a problem if the 3G wireless standard replaces the current 2.5G and if smart homes proliferate.6

Configuration A client running the IPv6 protocol can automatically configure itself with a unique address, eliminating the need for static addresses or previous methods of auto configuration such as DHCP (Dynamic Host Configuration Protocol).

The management of multiple IPv4 clients within an organization involves tracking the assignment of addresses either for each client, or for “pools” of clients.

Data Delivery There are new header fields in IPv6, which indicate the type of information being sent within each packet. This information can be used to prioritize traffic and guarantee Quality of Service (QoS)7. However, it is important to note that the actual implementation of QoS is still in the “research and development” stage as IPv6 alone is not sufficient for implementing end-to-end QoS.

For the transmission of multimedia data over the Internet, the fast and reliable delivery of IP packets is critical. Prioritization is one method of increasing speed and interactivity within the existing network topologies.

Routing IPv6 packets are moved from segment to segment using a simplified, hierarchical routing structure.

Routing under IPv4 is only partially hierarchical, relying also on large flat routing tables that can exceed 70,000 entries. Routing under IPv6, with its significantly smaller routing tables, requires less overhead at the router and is therefore more efficient and faster.

5 In Pakistan, a class C address in 2000 cost between $1050 and $1275 a year . Due to a lack of addresses, the price of a class C address almost doubled. By 2002, a class C address cost between $1900 and $2300 a year. 6 “'Smart' Homes for Smart People,” Wired News [online], 1999, http://www.wired.com/news/business/0,1367,17676,00.html [accessed 3/27/2004]. 7 Suydam, M. “Blazing trails: By paving paths for packets, MPLS could clear the way for IP convergence,”– CommVerge [online], 2002 http://www.reed-electronics.com/ednmag/index.asp?layout=article&articleid= CA214592&rid=0&rme=0&cfd=1 [accessed 3/27/2004].



Security IP security standards (IPSec), previously optional under IPv4, are now required under IPv6.

Standardized, layer 3 security reduces hacking activities.

Mobile Current implementation of mobile IPv4 requires the use of a foreign agent (FA), a home agent (HA), and a care-of (CO) address. The FA has to communicate the CO address through a tunnel back to the HA on the user's home network. The packets from the corresponding node to the mobile unit always have to go through the HA. IPv6 uses similar but more efficient process. The auto-configuration feature of IPv6 enables the mobile nodes to configure its own address without the help of any servers other than a router. Route optimization signaling enables a mobile IPv6 node to inform its correspondent node about its new care-of address. This allows both mobile node and the correspondent node to send and receive packets using the shortest path between the two.

No special mechanism is necessary on organization’s networks to support Mobile IPv6, other than home agent (embedded in IPv6 protocol). The large address space ensures that the auto-configured address on the mobile node does not conflict with the existing addresses of the network. Resulting in ubiquitous support for mobile Internet access and increase support of wireless devices such as PDAs and Pocket PCs by ISPs.

Multicasting The built-in multicasting in IPv6 allows a server to send a single packet with multiple addresses. The ISP will do the final routing. This reduces the bandwidth required for multimedia applications and broadcasting.

Allows several levels of multicasting and the creation of routing trees. This is a more efficient routing mechanism for applications such as Jini, which depend upon the ability to “discover” compatible devices on the network. Similar mechanism is used in Universal Plug and Play. Also, improve the distribution of multi media applications such as video steaming.

(Adapted from Microsoft, 2000)8

8 Microsoft Corporation. Introduction to IP version 6 [online], 2003.

http://download.microsoft.com/download/5/2/5/525343cc-7ba4-4e3b-a96a-c7a040d98d2d/IPv6.doc, [accessed March 27, 2004].



Appendix B: Procedure used to search trade journals

To compare the amount of information about IPv6 that is available to managers,

we compared the number of articles on that topic to the number of articles on two other

standards that have been introduced at approximately the same time. The search was

restricted to years 1998 through 2004. The following describes the procedure used to

complete the search.

Four top trade publications were searched for the number of articles published on

three topics: IPv6, Windows XP, and XML. Those publications were:

1. InfoWorld 2. Information week 3. Computer world 4. CIO magazine

The search was conducted by visiting the web sites of all four publications.

However, the sites did not provide equivalent search facilities. For example, all sites

featured an “advanced search,” however the search engine of InfoWorld magazine

behaved erratically and displayed different results every time a search was repeated,

therefore the search was conducted six times for each technology and the results were

then consolidated after eliminating duplicate entries. To triangulate the search results, the

ProQuest and ABI/Inform bibliographic database were used

For keywords that returned excessive number of matches, the search was

restricted to article titles only. The search results were then separated by year.

The Changing Dynamic of the Internet

Documents

Transcript of The Changing Dynamic of the Internet