TY - JOUR
T1 - Performance prediction for supporting mobile applications’ offloading
AU - da Silva Pinheiro, Thiago Felipe
AU - Silva, Francisco Airton
AU - Fé, Iure
AU - Kosta, Sokol
AU - Maciel, Paulo Martins
PY - 2018/5/23
Y1 - 2018/5/23
N2 - Mobile cloud computing (MCC) is a technique for increasing the performance of mobile apps and reducing their energy consumption through code and data offloading. Developers may build MCC applications on a public cloud. The public cloud may offer economies of scale, but there are some considerations to take into account. Cloud providers charge their customers by data traffic, and wrong offloading decisions may lead to financial losses. This paper proposes an approach for estimating applications’ performance, data traffic generated by tasks offloading and its related costs on a public cloud. This work proposes both a stochastic Petri net (SPN)-based formal framework to represent MCC applications in a method-call level and a cost model to predict data traffic volume. Our approach enables designers to plan and tune MCC architectures based on three performance metrics: mean time to execute, cumulative distribution function, and throughput. Our SPN-based framework represents the use and sharing of the bandwidth available for offloading operations as well as the effect of bandwidth variation on the evaluated metrics. It allows a more accurate evaluation by developers about the performance of their applications taking into account specific network requirements, users, and offloading strategies. Two case studies were performed. Our approach has proven to be feasible, and it highlights the most appropriate strategies, supporting developers at design time by providing statistical information about applications’ behavior and costs estimations.
AB - Mobile cloud computing (MCC) is a technique for increasing the performance of mobile apps and reducing their energy consumption through code and data offloading. Developers may build MCC applications on a public cloud. The public cloud may offer economies of scale, but there are some considerations to take into account. Cloud providers charge their customers by data traffic, and wrong offloading decisions may lead to financial losses. This paper proposes an approach for estimating applications’ performance, data traffic generated by tasks offloading and its related costs on a public cloud. This work proposes both a stochastic Petri net (SPN)-based formal framework to represent MCC applications in a method-call level and a cost model to predict data traffic volume. Our approach enables designers to plan and tune MCC architectures based on three performance metrics: mean time to execute, cumulative distribution function, and throughput. Our SPN-based framework represents the use and sharing of the bandwidth available for offloading operations as well as the effect of bandwidth variation on the evaluated metrics. It allows a more accurate evaluation by developers about the performance of their applications taking into account specific network requirements, users, and offloading strategies. Two case studies were performed. Our approach has proven to be feasible, and it highlights the most appropriate strategies, supporting developers at design time by providing statistical information about applications’ behavior and costs estimations.
KW - CTMC
KW - Data traffic evaluation
KW - Mobile cloud
KW - Performance evaluation
KW - Stochastic Petri nets
UR - http://www.scopus.com/inward/record.url?scp=85047273253&partnerID=8YFLogxK
U2 - 10.1007/s11227-018-2414-6
DO - 10.1007/s11227-018-2414-6
M3 - Journal article
AN - SCOPUS:85047273253
SN - 0920-8542
VL - 74
SP - 4060
EP - 4103
JO - Journal of Supercomputing
JF - Journal of Supercomputing
IS - 8
ER -