Tom Vest

Background: First Internet experience (Fort Lee DoD mainframe / Adventure c. 1979)

Background: Undergrad UVa/band/ skiing/languages/philosophy/architecture

Background: Taiwan residency 1990-1991

Background: China / Cui Jian / Identity Crisis 1991

In a band in China, played on TV

Background: Brief State Department career bid 1992-1993

Background: Grad school @ USC

-- Why telecom policy?

-- Alker / Santa Fe Institute

-- Mayor Riordan / Los Angeles Ofc. International Trade

-- Abe Lowenthal / Pacific Council story

-- Meso-level regional-international connections, including trade, immigration, investment, and Hollywood (media, audiovisual entertainment)

-- APEC / PECC Tel

-- FCC International Bureau

WoN: Cost Paradigm shift for composite service delivery

-- Asymmetrical competition, openness -> arbitrage

-- US domestic context: UNE

-- ITU / international telecom context: Benchmark Settlements

* get facts and dates *

-- US Internet context: "Great Depeering" in 1997

* get facts and dates *

-- International Internet context: ICAIS

WoN: IP peering

-- Relationship between infrastructure and cost of service

-- Why not just traffic volumes/balancing?

-- Pre-Netflow (& PCH work in this area)

-- Traffic too malleable (Norton's "Peering Playbook")

Question: How could an ISP have traffic without infrastructure?

-- Insourcing vs. outsourcing (more below)

Question: Did actual costs change, or was the paradigm shift more subtle?

Question: How to solve to the settlement crisis?

Question: How does the new paradigm relate to ICAIS?

Background: hired by AOL 2000

At AOL: Originally charged with peering with foreign ISPs/telcos (connection with PhD studies)

At AOL: Thin international presence + limited interest in international -> broad responsibilities to provision all components

WoN: Changing costs: Dynamics of insourcing vs. outsourcing network infrastructure

WoN: IP transit vs. infrastructure == retail (outsourcing) vs. wholesale (insourcing)

WoN: TInker toy methodology for provisioning subsequent network deployments

At AOL: In the end, duties included analysis, pricing, scenario design and approval, provisioning and logistics, field deployment, peering and relationships with CN, JP. AU joint venture partners -- overall optimization of the end-to-end deployment equation

At AOL: New capex justified against how quickly investment would be repaid in opex savings; dynamic contributed to AOL's ascent up the peering hierarchy. AOL now self-identifies ATDN as "Tier One network"

WoN: All Tier 1 ISPs that are not legacy IXCs primarily composed primarily of such tinkertoy components; even post-IXC Tier 1s encompass large portions assembled this way.

Background: Joined PCH in Spring 2004, known Bill for several years; worked on projects that leveraged knowledge of network economics, and experience with US, international policy makers. PCH primary mission building IXPs, gross mechanisms for multiplexing circuits, hardware. Started writing book/dissertation again.

WoN: Multiplexing as "secret sauce of the Internet" -- technical divisibility of all critical inputs required for end-to-end provisioning. Statmux is just one example -- the benefit that packet switching provides over circuit switching within a given pit-to-point link. Other examples include optical, hardware (e.g., IXP), site (e.g., neutral IDCs), etc., etc. When combined with the right institutional arrangements, these fractional components can be enjoyed independently by different institutions, with the only collateral effects being competition.

WoN: Put it all together == "logical multiplexing" -- the ability of different institutions representing different "visions of the good" (i.e., strategies for delivering some valuable service to some important constituency at an overall cost that can be internally justified) to independently leverage a shared set of basic infrastructure components, thereby optimizing overall efficiency of infrastructure utilization in the delivery of more and "better" Internet services over time.

Background: Accepted invitation to join CAIDA part-time in Feb 2006; KC was one of the people I approached early about WoN concept. Continuing relationship with PCH.

Question: Whither the "bandwidth glut?"

WoN: The possibility of a new discipline of network economics

-- Origins of the discipline of macroeconomics

Tom Vest

Attachments