Article Preview
TopIntroduction
Cloud Computing is popular resource platform where user can offload their applications for executing and get result back in order to overcome the limitations of mobile device. Mobile Cloud Computing is emerged by a combination of two popular technologies such as Cloud computing and Network communications. MCC has been gained a lot of attention from researchers. The most advantage of MCC is that it reduces the complexity of the application and improves mobile device performance in terms of power. Mobile Devices (MD) is becoming more powerful for running complex applications such as resource intensive applications. The limitations of mobile devices in terms of battery, CPU speed, and limited memory are making developers unable to run the complex applications (Devare et al., 2010).
In order to improve the performance of the Mobile device, MCC has introduced a Novel concept called Offloading, which can offload resource intensive application into the Cloud. There are various Cloud service providers such as icloud and EC2. Mobile users can use resources of elastic Cloud in order to optimize performance of mobile applications. In (Kumar & Lu, 2010), the author has focused on energy utilization of mobile device, by offloading tasks into Cloud environment to improve the performance of the mobile device. In (Z. Li et al., 2001)(Rong & Pedram, 2003), the author offloading computation task to remote Cloud to reduce energy utilization. The networking cost between mobile devices and remote servers was addressed in (Gu et al., 2003) . The application is portioned and offloaded to the nearby remote server for processing(Krishna et al., 2016). The response time between mobile device and Remote Cloud is significant challenge. However, the algorithm has not addressed response time in both wireless and remote environment(Raju & Saritha, 2016). The MCC is used in various areas such as image processing, Speech recognition, Translator etc.(Dinh et al., 2013)(Gkatzikis & Koutsopoulos, 2014)(Y. Wang et al., 2015) (Rahimi et al., 2014)(Sheikhalishahi et al., 2011).
The mobile device can run high end applications which require huge computation power and Storage. The requirement of resources for each application may vary. In result, Resource allocation to mobile devices should be dynamic. In MCC, mobile devices can offload intense application to the remote Cloud for faster execution(Chun et al., 2011)(Cuervo et al., 2010)(Gkatzikis & Koutsopoulos, 2013)(R. Somula et al., 2019). The distance between mobile device and Cloud remote server increases the response time. In result, the overall execution time of intensive application also increases the new concept called Cloudlet has been introduced to address latency related challenge (Satyanarayanan et al., 2009). Cloudlet is small scale data center which is available around the user. By using nearby Cloudlet for execution, the user can decrease the overall response time of application. Cloudlet is a three-tier architecture in Fig.1, which is introduced between mobile devices and remote Cloud. The virtual technology available in Cloudlet to share hardware resources with incoming requests from users. The resources of Cloudlet (available bandwidth, CPU, Memory, etc.) are shared with VMs. In Cloud computing resource processing is a popular area (Bobroff et al., 2007)(Das et al., 2013)(Das et al., 2014)(Jiwei Li et al., 2013)(Van et al., 2009). Resource scheduling among Cloudlets is a significant issue(Gkatzikis & Koutsopoulos, 2014). The mobile user always moves from one location to another location. Therefore, the distance also increases between mobile device and Cloudlet; it causes a delay in execution time. In order to address this issue, the task is moved to nearest Cloudlet by measuring user distance. The distance between mobile device and Cloudlet is not the only reason for task migration to another Cloudlet. The load of the target Cloudlet also one important factor for task migration, when the server handles a greater number of tasks than actual capacity then execution time of task increases gradually. Some works have focused on VM scheduling in MCC (Gkatzikis & Koutsopoulos, 2014)(L. Wang et al., 2014)(Islam et al., 2016)(Liu et al., 2015)(Taleb & Ksentini, 2013)(R. Somula & Sasikala, 2019)(Ramasubbareddy & Sasikala, 2019). The previous works focused on static task execution. In this proposed method, we consider bandwidth, load and distance as constraints to select Cloudlet for scheduling among Cloudlets. In this paper, the proposed model addresses the scheduling among Cloudlets. Therefore, resource constrained devices show better performance. The novelty in this paper is considering three objective functions to schedule VM in another Cloudlet (Distance, Bandwidth, Memory and Load of the Cloudlet).