Risks in Information Networks and Technology: The insidious decay of IT infrastructure

This presentation argues that Information Technology is infrastructure and should be designed accordingly. The talk draws an analogy between the computer/network system and the railroad — as the railroad transformed industrial society, the network-and-computer system transforms the information economy, with comparable requirements for durability, interoperability, and institutional accountability. The talk introduces the decay/disaster distinction: software does not wear out, but it decays — through brittle architecture exposed by component upgrades, mission creep, maintenance-induced bugs, and "bit rot" (the loss of information through loss of the means to read it). Decay is slow and insidious; disaster is fast and obvious. The dangerous failure mode is the slow one. Design principles proposed: use open formats and protocols; prefer boring, proven solutions over proprietary or bleeding-edge technology; require multiple sources; seek graceful degradation; maintain local human resources as the only reliable first response. A section addresses the specific vulnerabilities of Latin American and Caribbean contexts, where economic instability compounds technical decay. The final section discusses Peru's open-source bill (Congressman Edgar Villanueva, 2002), whose three principles — no single provider encodes data; no single provider controls usability and maintainability; transparency and accountability — are shown to lead independently to robust, decay-resistant, recovery-friendly IT. This is an early documented statement of the data-sovereignty argument as an architectural principle rather than a regulatory claim.Talk delivered at the Second Hemispheric Conference on Vulnerability Reduction of Trade Corridors to Socio-Natural Disasters (TCC II), Tampa, June 13, 2002.