Mobile Cloud Computing Architecture for Computation Offloading

Archana Kero; Abhirup Khanna

Outline

Mobile Cloud Computing Architecture for Computation Offloading

Archana Kero

Abhirup Khanna

Abstract

— In recent years the Smartphone has undergone significant technological advancements but still remains a low computational entity. Mobile Cloud Computing addresses this problem and provides a solution in form of Mobile Computation Offloading (MCO). Computation offloading is a concept in which certain parts (tasks) of an application are executed on cloud whereas the rest of them on the mobile device itself. MCO turns out to be a great help with respect to resource constrained mobile derives as it allows resource intensive tasks to be executed remotely. Though such procedures save mobile resources but incur communication cost between the mobile device and cloud. Thus, it becomes extremely essential for offloading models to take steps (offloading decisions) in order to augment the capabilities of mobile devices along with reduced execution and communication costs. In this paper, we present an application offloading model which offloads an application based upon the nature and execution pattern of its tasks. We also propose an algorithm that depicts the work flow of our computation model. The proposed model is simulated using the CloudSim simulator. To this end, we illustrate the working of our proposed system along with the simulated results.

FAQs

What are the key components of the proposed mobile computation offloading model?add

The proposed model includes four key components: Application Analyzer, Network Profiler, Decision Engine, and device profiler, each coordinating to optimize task execution.

How does the proposed model categorize application tasks for offloading?add

Tasks are categorized as either CPU intensive or I/O intensive, with the model primarily focusing on offloading CPU intensive tasks based on execution parameters.

What role do communication and execution cost functions play in the model?add

The communication cost function estimates battery consumption and time for tasks, while the execution cost function quantifies the execution efficiency based on device capacity and task complexity.

What was the effect of increasing offloaded tasks on execution time?add

The study observed execution time decreasing from 9000 to 4400 seconds as the number of offloaded tasks increased, with a rise in time occurring past ten tasks.

How does the model determine the feasibility of computation offloading?add

Feasibility is assessed by comparing the Final Resultant of local execution against that of offloaded tasks; offloading is deemed beneficial if the resultant from offloading is lower.

Figures (6)

The above proposed mathematical model showcases various equations which talk about the entire process of computation offloading from categorizing a particular task to its execution at the cloud end. We make use of a Directed Acyclic Graph to represent the entire workflow of task execution. The following figure corresponds to one such DAG wherein task execution takes place. Fig. 1. DAG for Task Execution Graph

Fig. 2. Offloading Model In this subsection, we would be talking about our application model of our proposed system. The model comprises of four actors which communicate among themselves and describe the working of our offloading system. The model is well distributed across multiple computing environments and thus has some of its components running on Cloud whereas the others on the mobile device itself. Following is a block diagram which illustrates the application model.

Fig. 3. CloudSim Simulation Following is the parameter table that exhibits all the various parameters that have been considered during simulation.

Fig. 5. Experiment Result The above graph showcases the variation in total execution time of an application in seconds. The X axis represents the number of offloaded tasks whereas the Y axis represents the execution time. The graph clearly depicts how the total execution time of an application started to reduce from 9000 to 5500 and finally 4400 seconds by increasing the number of offloaded tasks. But a sudden rise in execution time can be seen once the number of offloaded tasks were increased from 6 to 10. This experiment makes it very much evident that offloading does reduce execution time and is beneficial in most of the cases but can even turn out to be expensive if applied extensively.

Fig. 4. Experiment Result Furthermore we performed an experiment wherein we steadily increased the number of tasks being offloaded and studied the execution pattern of the respective application. Following is the graph which showcases variation in total execution time (execution time + communication time) of an application when its number of offloaded tasks kept on increasing.

References (19)

Input: An application for computation offloading Output: Total Execution Time & Battery Consumption
Device Profiler divides the application into numerous tasks (N)
Type of task T(V i ) = {I/O || CPU};
AZ → ei(L);
NP → ti(L);
CM(β, τ) = { ei(L) & ti(L)};
Total cost function is computed call (F(T(V)i, L));
DE → FR(T, (V i ), L);
compute (O) = { 0|| 1};
Khanna, A. (2015, September). RAS: A novel approach for dynamic resource allocation. In Next Generation Computing Technologies (NGCT), 2015 1st International Conference on (pp. 25-29). IEEE.
Enzai, N. I. M., & Tang, M. (2014, April). A taxonomy of computation offloading in mobile cloud computing. In Mobile Cloud Computing, Services, and Engineering (MobileCloud), 2014 2nd IEEE International Conference on(pp. 19-28). IEEE.
Kovachev, D., Yu, T., & Klamma, R. (2012, July). Adaptive computation offloading from mobile devices into the cloud. In 2012 IEEE 10th International Symposium on Parallel and Distributed Processing with Applications (pp. 784-791). IEEE.
Khanna, Abhirup, Sarishma. (2015). Mobile Cloud Computing: Principles and Paradigms. IK International.
Dinh, H. T., Lee, C., Niyato, D., & Wang, P. (2013). A survey of mobile cloud computing: architecture, applications, and approaches. Wireless communications and mobile computing, 13(18), 1587-1611.
Calheiros, R. N., Ranjan, R., Beloglazov, A., De Rose, C. A., & Buyya, R. (2011). CloudSim: a toolkit for modeling and simulation of cloud computing environments and evaluation of resource provisioning algorithms. Software: Practice and Experience, 41(1), 23-50.
Jiang, Y., He, J., Li, Q., & Xiao, X. (2014, December). A dynamic execution offloading model for efficient mobile cloud computing. In 2014 IEEE Global Communications Conference (pp. 2302-2307). IEEE.
Cheng, Z., Li, P., Wang, J., & Guo, S. (2015). Just-in-time code offloading for wearable computing. IEEE Transactions on Emerging Topics in Computing, 3(1), 74-83.
Shiraz, M., Sookhak, M., Gani, A., & Shah, S. A. A. (2015). A study on the critical analysis of computational offloading frameworks for mobile cloud computing. Journal of Network and Computer Applications, 47, 47-60. 2016 2nd International Conference on Next Generation Computing Technologies (NGCT-2016)
Dehradun, India 14-16 October 2016

Mobile Cloud Computing Architecture for Computation Offloading

Sign up for access to the world's latest research

Abstract

FAQs

Related papers

References (19)

Related papers