Fault-tolerance in Delta-4

Fault-tolerance and distribution go hand in hand. Should a single element of a distributed system fail, users expect at worst a slight degradation of the service that they are offered; distributed systems must thus at least have some built-in fault-tolerance. Preferably, distribution allows redundant resources to be interconnected and used in such a way as to provide continued undegraded service in the presence of faults. The Delta-4 project pursues this tight relationship between distribution and fault-tolerance in order to offer distributed architectures with built-in fault-tolerance features. Applications supported by Delta-4 architectures can be made incrementally fault-tolerant on a service-by-service basis. At application configuration time, the application designer can choose which services he wishes to make fault-tolerant and to which degree. Several techniques for fault-tolelrance are supported; the application designer can thus choose, according to the available resources, the technique that best suits his application and makes the best trade-off between cost, fault-tolerance and performance.