Resource Usage Prediction Models for Optimal Multimedia Content Provision
Kryftis, Y. K.
; Mavromoustakis, C. M.
; Batalla, J. B.
Rodrigues, J. R.
; Dobre, C. D.
IEEE Systems Journal Vol. 4, Nº 11, pp. 1 - 12, December, 2017.
ISSN (print): 1932-8184
Journal Impact Factor: 0,923 (in 2011)
Digital Object Identifier: 10.1109/JSYST.2016.2548423
Download Full text PDF ( 1 MB)
This paper proposes a network architecture that uti- lizes novel resource prediction models for optimal selection of multimedia content provision methods. The proposed research approach is based on a prototype system, which exploits a resource prediction engine (RPE), utilizing time series and epidemic spread models, for optimal and balanced distribution of the stream- ing data among content delivery networks, cloud-based providers and home media gateways. The proposed epidemic diseases mod- els adopt the characteristics of the multimedia content delivery over the network architecture. In this context, this paper aims to present the advantages of using such models, by presenting and analyzing an epidemic spread scheme for video-on-demand (VoD) delivery, to predict future epidemic spread behavior. In addition, this paper presents two algorithms, adopted in the pro- totype network architecture, for optimal selection of multimedia content delivery methods, as well as balanced delivery load, by exploiting the RPE. Both algorithms and models are evaluated to establish their efficiency, toward effectively predicting future net- work traffic demands. The simulation results verify the validity of the proposed approach, identifying fields for future research and experimentation.