Protocols for End-To-End Reliability in Multi-Tier Systems

Internet services exhibit the trend to be structured according to a three-tier, and in general multi-tier, system organization. Even though the partitioning of applications into multiple tiers provides the potentialities to achieve high modularity and ¿exibility, the multiplicity and diversity of the employed components, and their interdependencies, make reliability a complex issue to tackle. In this book we present protocols ensuring the e-Transaction (exactly-once Transaction) guarantees, a recent formalization of end-to-end reliability properties for multi-tier systems. The presented solutions achieve high scalability and exhibit very limited overhead, revealing particularly attractive for large scale service delivery platforms. From a theoretical standpoint, the presented e-Transaction protocols can cope with purely asynchronous systems, where no assumption on the failure detection accuracy can be guaranteed. Some of the building blocks underlying these fault tolerant protocols can also be used to deal with a more general class of failures, which we refer to as "performance failures", which model situations of reduced system responsiveness due to both crashes and overloads.