COUFLESS 2015 - 1st International Workshop on Complex faUlts and Failures in LargE Software Systems

Big data is becoming an integral part of modern large software systems. Many organizations and companies deploy large-scale distributed applications which use big data by sending requests to data management components (i.e., database management systems and data storage management layers) and obtaining data that result from the execution of these requests. Mandelbugs in these applications occur as a result of conditions that accrue within the databases and a significant time lag between fault activation and failure is due to error states that cross between the components of these applications and the databases. The term Mandelbug refers to faults whose behavior seems to be “non-deterministic”. It is difficult to isolate and the failures caused by it are hard to reproduce. Mandelbug characteristics have been connected with the complexity of its activation or error propagation.
Performance is a key challenge in building large distributed applications since their predictability is inherently difficult due to Mandelbugs, complex faults and failures that require a uniform set of models that link the behavior of applications with data constraints and the properties of data management software.