In a decentralised system the problems of fault tolerance, and in particular error recovery, vary greatly depending on the design assumptions. For example, in a distributed database system, if one disregards the possibility of undetected invalid inputs or outputs, the errors that have to be recovered from will just affect the database, and backward error recovery will be feasible and should suffice. Such a system is typically supporting a set of activities that are competing for access to a shared database, but which are otherwise essentially independent of each other--in such circumstances conventional database transaction processing and distributed protocols enable backward recovery to be provided very effectively. But in more general systems the multiple activities will often not simply be competing against each other, but rather will at times be attempting to co-operate with each other, in pursuit of some common goal. Moreover, the activities in decentralised systems typically involve not just computers, but also external entities that are not capable of backward error recovery. Such additional complications make the task of error recovery more challenging, and indeed more interesting. This paper provides a brief analysis of the consequences of various such complications, and outlines some recent work on advanced error recovery techniques that they have motivated.

Exception handling is not only useful for increasing program readability, but also provides an effective means to check and locate errors, so it increases productivity in large-scale program development. Some typical and frequent program errors, such as out-of-range indexing, null dereferencing, and narrowing violations, cause exceptions that are otherwise unlikely to be caught. Moreover, the absence of a catcher for exceptions thrown by API procedures also causes uncaught exceptions. This paper discusses how the exception handling mechanism should be supported by the compiler together with the operating system and debugging facilities. This mechanism is implemented in the compiler by inserting inline check code and accompanying propagation code. One drawback to this approach is the runtime overhead imposed by the inline check code, which should therefore be optimized. However, there has been little discussion of appropriate optimization techniques and efficiency in the literature. Therefore, a new solution is proposed that formulates the optimization problem as a common assertion elimination (CAE). Assertions consist of check code and useful branch conditions. The latter are effective to remove redundant check code. The redundancy can be checked and removed precisely with a forward iterative data flow analysis. Even in performance-sensitive applications such as telecommunications software, figures obtained by a CHILL optimizing compiler indicate that CAE optimizes the code well enough to be competitive with check suppressed code.