Single point failures
From: David Lacey’s IT Security Blog | September 02, 2009
Indicates that the author believes that the use of Cloud Computing style technologies is introducing significant single points of failure. His evidence for this a failure in Google’s gmail. Surely such failures are not inevitable, as this technology is design to reduce such failures through multiple redundancy. The failures do happen, but I would propose they are not inevitable. They are more a symptom of pressures to circumvent due development process, thus reducing cost and time to market.
In my mind, the question really is whether Google accepted that there was a significant risk of system outage in order maintain reduced cost for Gmail delivery? If so, was this an appropriate choice for the market?