Context
Robust distributed computingTask Description
Computer may crash. At a single single site this means the end of all computations running there. Using a network of distributed computers, the crash of one site shall not cause the stop of a distributed computation. Moreover, it shall be considered which data have to survive a crash. These requirements shall be analysed using classical problems for distributed processing. We offer a toolkit that already offers data persistency and also the possibility to specify check points for computations and thus makes it easy to develop robust, distributed programs. The focus is to gain practical experience: what does recoverability mean for the single problem classes? Tests shall be done by using systematic software fault injection.
Requirements
- Erfahrung mit verteilter Programmierung, aber nicht notwendig.
Student
Open Topic
Supervisor
A.o. Univ.-Prof. Dr. Dipl.-Ing. eva Kühn (please contact eva Kühn)
