A Tool for the Expression of Failure Detection Protocols

Failure detection protocols - an important building block of fault-tolerant distributed systems - are often described by their authors making use of informal pseudo-codes of their own devising. Often these pseudo-codes use syntactical constructs that are not available in COTS programming languages such as C or C++. This translates into informal descriptions that require ad hoc interpretations and implementations. Being informal, these descriptions cannot be tested by their authors, which may translate into insufficiently detailed or even faulty specifications. Being non-standard, they require the reader to understand each time a different lingo. This paper tackles the above problem introducing a formal syntax for the expression of failure detection protocols and a C library that implements that syntax: a toolset to express and reason about failure detection protocols. The resulting specifications are more diffuse but non ambiguous and eligible for becoming a standard form among researchers and practitioners