George A. Barnett, International Telecommunication Networks, 24 October 2011
From Driscollwiki
George A. Barnett UC Davis 24 Oct 2011
International telephone network
- 31 points in time 1978-2009
- Sources: ITU ("TeleGeography"), AT&T ("World's Telephones")
- Network includes 164 countries, very dense
Centrality by rank order
- Fits a "decaying exponential"
International telephone network, 2009
- Minimum of 1m minutes of call time to make an edge
- Uses colors to ID regions
Results support
- World Systems Theory (Barnett)
- Galtung's structural theory
- Huntington's notion of civilizations (language, culture, religion)
Centrality over time
- Some shifts, smaller European countries fall as Asian countries rise
Animation of network using ET-V
- Developed with Benjamin Elber
Growth
- Explained by immigrants (Philippines, India)
- Adoption of mobile telephones
- In many cases, calls originate with mobile, terminate at landlines
- Asymmetrical adoption of mobile
Decline in intl telephone in the US
- Partially explained by email, IM, etc.
- Adoption of VoIP
- Greatest growth in India, Pakistan, Bangladesh
- "Today, Skype is the largest provider of cross-border voice communications"
- Meanwhile, residential lines decline
Next steps
- End of intl telephone comm
Criticism
- International Telephone Network (POT), not all intl telecomm
- Non directional only (Monge & Matei, 2004)
- Ahn & Barnett (1995) examined intl telex network
- Salisbury & Barnett Salisbury intl Visa network
- OECD hyperlink
International internet hyperlink
- Barnett & Park (@005)
- Using data from alta vista (jan31, 2003)
- 550m websites, 356m links
- Primarily English Lang sites
- Simply algo: looking for connects between .ca and .uk
- US: .edu .mil. gov .us
- USA may be more or less central
Intl hyperlink reliability
Sources of invalidity
- Overlapping uses of TLD: ".edu.au", ".co.uk", ".gov.co" (Colorado)
International internet bandwidth
- Bilateral bandwidth capacity from TeleGeography (2003)
International internet infrastructure
- By design, lots of traffic flowing across US
- e.g. strongest ties from Asia to Europe flows across US
Non-economic factors
- Language (Barnett & Choi, 1996)
- History (J. Kim * Barnett, 2007)
Across studies of many international flows
- Similar patterns
Current research
- Replicated Barnett & Park (2005)
- International Internet hyperlink data, May, 2009 + 2010
- Yahoo API changed and he can't use the same method any longer
Decomposing (cracking) gTLDs (.com .org .net)
- Paper in Social Science Computer Research and Evaluation
Barnett's "why"
- Not tech but social/comm theory
- Structure of global community
- Changes over time: globalization
Cultural convergence theory (Rogers & Kincaid, 1981)
- "Participants will converge on an average pattern of thought (global culture) if communication is unrestricted"
- Reformulated by Barnett and someone in 2007
Regional groupings with greater concentrations of communication...
Q&A
What might happen if you remove the US from the network?
- Possibly locate more regionalization
- But may be less "true" because 20% of the calls orig/term in US
To which nodes does the Latin American group connect?
- Spain, Portugal
- Spain is in an interesting position
- Portugal is an important Euro node but Brazil is much more central overall
- Not as strong ties with Japan as you might expect

