This seminar discusses extensions to the functional language Erlang which provide a secure execution environment for remotely sourced code. This is in contrast to much existing work which has focused on securing procedural languages. Using a language such as Erlang provides a high degree of inherent run-time safety, which means effort can be focused on providing a suitable degree of system safety. The changes needed were providing unforgeable (capability) references with access rights to control the use of system resources; the provision of a hierarchy of execution nodes to provide custom views of the resources available and to impose utilisation limits; and support for remote module loading.