Past few years, Big Data and cloud computing have become buzzwords in IT
region, and we have been seeing that data are generated in massive amounts and at an
increasing rate in all domains. The reliability and efficiency of distributed systems
have always been a major concern of the service providers and users. Therefore, fault
tolerance is among the most essential issues in distributed clouds to deliver reliable
services to customers.
In Big Data domain, scientific workflows are increasingly used for Big Data
analysis, processing, and management. With movement the world to Big Data, singlesite processing becomes unsuitable and Big Data scientific workflows can no longer
be accommodated within a single computing system, and ensuring a level of reliability
for a scientific workflow execution is a complex task that will tend to increase the
cost.
Replication of tasks increases redundancy and thereby the reliability, which is
achieved by parallel execution of a task on multiple virtual machines simultaneously
to guarantee a viable result, which leads to a high cost.
This doctoral Thesis presents a fault-tolerant model with two approaches that
optimize the reliability and execution cost of Big Data scientific workflows on cloud
computing environments and ensure a predefined level of reliability by replicating
tasks.
Finally, the model was implemented using WorkflowSim, it is extension of the
CloudSim simulator framework that is used for modelling and simulation of cloud
computing infrastructures and services.
ISBN: | 978-80-7678-032-3 |
EAN: | 9788076780323 |
Počet stran |
46 stran |
Datum vydání |
19. 10. 2021 |
Pořadí vydání |
První |
Jazyk |
anglický |
Vazba |
e-kniha - pdf |
Autor: |
Ammar Nassan Alhaj Ali |
Nakladatelství |
Univerzita Tomáše Bati ve Zlíně |
Tématická skupina |
999 - nezařazeno |
| Neprodejná publikace. Publikaci je možné poptávat zde: Voně dostupné na http://hdl.handle.net/10563/50069 |