Recent Skype outage, apparently fault of a denial-of-service attack on the Skype centralized login infrastructure, raises again the intrinsic flaw of designing a service or an application (even partially) based on a centralized network topology.
As the recent facts show, offering a service with a Single Point of Failure creates a “domino effect” whose legal implications (in terms of damages suffered by paying clients) might bear unforeseen consequence. The “flawed-by-design” kind of liability might, indeed, lead to a class action against Skype for having knowingly chosen to build their service on a technical structural weakeness.
Of course, I imagine that should that issue be taken into Court, ICT expert witnesses will play the major role in addressing the underlying technical issues.