The hardware environment

Abstract

The paper considers how security protocols are as they are in part because of the hardware environment in which they are expected to function. Expectations were set nearly twenty years ago, when communications were very unreliable and slow, when reliable sources of time were most unusual, when encryption was extremely slow, when memories and disks were small and slow. In consequence protocol designers went to great lengths to minimise the number of messages sent and their size, particularly the size of the encrypted part. It was considered very undesirable to rely on time other than very locally for ordinal purposes, and systems were expected as far as possible to be stateless.