A Proposal to investigate the use of error correcting code techniques in implementing distributed systems resistant to Byzantine failures and security breaches

Throughout the literature on reliable distributed systems there is much coverage of systems which maintain correct operations in the face of fail-stop or non-Byzantine failures. What are less represented are methods for dealing with the harder problem of Byzantine failures. This paper proposes a method for dealing with these sort of failures. Fail-stop or non-Byzantine failures typically are characterised by the component or system involved ceasing to transmit or receive messages; effectively removing itself from the distributed system. Byzantine failures are typically ones where the component or system fails but continues to transmit invalid or incorrect messages causing failures or incorrect behaviour elsewhere in the system. Security breaches resulting in a system transmitting incorrect messages through malice rather than failure could also be viewed as a sort of Byzantine failure. In fact, by showing that a class of Byzantine failures and certain security breaches are equivalent, a solution of one will provide a solution for the other.