Leveraging Utilization as Performance Metric for CDN enabled Internet of Things

Published on Dec 1, 2019in Musculoskeletal science and practice
· DOI :10.1016/j.measurement.2019.07.042
Saif ul Islam2
Estimated H-index: 2
Hasan Ali Khattak5
Estimated H-index: 5
(CUI: COMSATS Institute of Information Technology)
+ 2 AuthorsJean-Marc Pierson19
Estimated H-index: 19
Abstract With the explosive increase in Internet users and ultimately its infrastructure, efficient resource utilization has become one of the most prominent challenges in large-scale distributed systems. Content Distribution Networks (CDNs) are among the favored large-scale distributed systems that are currently handling around 40 % of the Internet traffic. In literature, different approaches have been proposed to improve the CDN performance by improving CDN utility. The CDN utility is the ratio of the served contents to the pulled contents of a surrogate server. However, the utilization of surrogate servers has largely been neglected in the previous studies. Moreover, due to the vast expansion of such systems, their energy consumption has been tremendously increased. In fact, the energy consumption of the CDN networks is directly proportional to their utilization. Therefore, it is extremely important to consider the CDN utilization while designing and evaluating the performance of CDN policies. In this work we have considered a set of request redirection policies such as load balance and load unbalance. Furthermore, infrastructure size and volume of the client request traffic is also been taken into account. Our findings represent that even if surrogate servers are poorly utilized, the CDN utility exhibits higher values in general. These finding demonstrate that CDN utility does not take into account the utilization of surrogate servers. Even if the surrogate servers are poorly utilized, the CDN utility exhibits higher values. Conclusively, evaluating the CDN performance only on the basis of CDN utility without considering utilization of surrogate servers is an inefficient approach. Therefore, we propose to consider the CDN resources utilization as an important metric while analyzing the performance of a CDN.
  • References (23)
  • Citations (0)
#1Hafsa Arshad (CUI: COMSATS Institute of Information Technology)H-Index: 3
#2Hasan Ali Khattak (CUI: COMSATS Institute of Information Technology)H-Index: 5
Last.Samee U. Khan (NDSU: North Dakota State University)H-Index: 47
view all 5 authors...
#1Yixuan Wang (NUPT: Nanjing University of Posts and Telecommunications)H-Index: 2
#2Kun Wang (PolyU: Hong Kong Polytechnic University)H-Index: 28
Last.Song Guo (PolyU: Hong Kong Polytechnic University)H-Index: 39
view all 5 authors...
#1Tommaso Addabbo (UNISI: University of Siena)H-Index: 12
#2Ada Fort (UNISI: University of Siena)H-Index: 2
Last.Valerio Vignoli (UNISI: University of Siena)H-Index: 2
view all 6 authors...
#1Amir Hossein Alavi (MU: University of Missouri)H-Index: 39
#2Pengcheng Jiao (UPenn: University of Pennsylvania)H-Index: 11
Last.Nizar Lajnef (MSU: Michigan State University)H-Index: 17
view all 4 authors...
#1Javed Iqbal (Sarhad University of Science & IT, Ring Road (Hayatabad Link) Peshawar)H-Index: 4
#2Murad Khan (Sarhad University of Science & IT, Ring Road (Hayatabad Link) Peshawar)H-Index: 12
Last.Hasan Ali Khattak (CUI: COMSATS Institute of Information Technology)H-Index: 5
view all 7 authors...
#1Eugene Siow (University of Southampton)H-Index: 2
#2Thanassis Tiropanis (University of Southampton)H-Index: 12
Last.Wendy Hall (University of Southampton)H-Index: 40
view all 3 authors...
Cited By0
View next paperStudy on a CDN Performance on CERNET