Prediction-based prefetch scheduling in mobile service applications

Publikationstyp
Conference Paper
Date Issued
2016-12
Sprache
English
Author(s)
Borkowski, Michael  
Skarlat, Olena  
Schulte, Stefan  
Dustdar, Schahram  
TORE-URI
http://hdl.handle.net/11420/11873
Start Page
41
End Page
48
Article Number
7787053
Citation
IEEE International Conference on Mobile Services (MS 2016)
Contribution to Conference
IEEE International Conference on Mobile Services, MS 2016  
Publisher DOI
10.1109/MobServ.2016.17
Scopus ID
2-s2.0-85010471692
The usage of Web or Cloud-based applications on mobile devices suffers from varying link quality, which causes user-perceivable delays and therefore reduces the Quality of Experience (QoE). On the other hand, mobile devices are increasingly feature-rich, allowing to make usage of context data in order to predict network quality based on the user's location. In this paper, we propose a prefetch scheduling algorithm based on network quality predictions, and evaluate it using data collected from real-world field tests. We show that our approach fulfills the expected gain in QoE. Using network quality prediction to optimize data prefetching can improve the user-perceived response time by up to 95%.Our results not only show the feasibility of the proposed algorithm, but also motivate further research in the field of mobility pattern creation, and undermine the importance of location as a part of user context throughout the software stack.
Subjects
Connectivity
Location dependent services
Mobile networks
Prediction
Prefetching