Information Science

 In his book Tubes, Andrew Blum investigates the discrepancy in the presentation of computer networking as a cloud and the reality of its firm roots in infrastructure. This concept of belief that we are storing and using applications on a cloud can create an impression that services are more distributed and resilient to failure and breakdown than they are in reality. This is because data centres are very concentrated infrastructures and so are major switching hubs. This project is to define a recommended numerical degree of fully redundant infrastructure upon which it would be safe to describe the services as ‘cloud’. This can be achieved through the examination of scenarios and network failures based on real running infrastructure and assessing the associated damage to logical ‘cloud’ services running on top of these. Anticipated skills/knowledge required: A good understanding of global networks and infrastructures will be required, as will a good imagination and methodical planning to map potential outages required to demonstrate true resilience.