Elsevier

Future Generation Computer Systems

Volume 90, January 2019, Pages 149-157
Future Generation Computer Systems

Autonomic computation offloading in mobile edge for IoT applications

https://doi.org/10.1016/j.future.2018.07.050Get rights and content

Highlights

  • An autonomic computation offloading model for mobile edge/fog is proposed.

  • A deep reinforcement Q-learning model is used for computation offloading.

  • Our method significantly improves the performance of the computation offloading.

Abstract

Computation offloading is a protuberant elucidation for the resource-constrained mobile devices to accomplish the process demands high computation capability. The mobile cloud is the well-known existing offloading platform, which usually far-end network solution, to leverage computation of the resource-constrained mobile devices. Because of the far-end network solution, the user devices experience higher latency or network delay, which negatively affects the real-time mobile Internet of things (IoT) applications. Therefore, this paper proposed near-end network solution of computation offloading in mobile edge/fog. The mobility, heterogeneity and geographical distribution mobile devices through several challenges in computation offloading in mobile edge/fog. However, for handling the computation resource demand from the massive mobile devices, a deep Q-learning based autonomic management framework is proposed. The distributed edge/fog network controller (FNC) scavenging the available edge/fog resources i.e. processing, memory, network to enable edge/fog computation service. The randomness in the availability of resources and numerous options for allocating those resources for offloading computation fits the problem appropriate for modeling through Markov decision process (MDP) and solution through reinforcement learning. The proposed model is simulated through MATLAB considering oscillated resource demands and mobility of end user devices. The proposed autonomic deep Q-learning based method significantly improves the performance of the computation offloading through minimizing the latency of service computing. The total power consumption due to different offloading decisions is also studied for comparative study purpose which shows the proposed approach as energy efficient with respect to the state-of-the-art computation offloading solutions.

Introduction

The massive growth of mobile devices (e.g. smart phones, laptops, tablet pc’s, mobile IoT’s and automobiles) and their computation demands imposed a huge scarcity in communication network and computation resources. Some of the application services e.g. image processing and real-time translation services require extensive computation, the resource-constrained mobile devices are not the feasible domiciles to process those applications. Therefore, to meet the computation demands of such type of mobile devices and applications the outsourcing of computation is the demand in need.

Computation offloading is a relocation mechanism of processes or modules of software applications or systems from resource-constrained devices to the resource-rich platforms. Mobile cloud is the well-known platform for computation offloading of mobile devices. Mobile cloud computing is becoming a popular method for mobile services e.g. mobile video games, video streaming, education, social networking, messenger and mobile healthcare services [1].

However, the key barriers to offloading computation in mobile cloud are the network bandwidth and latency. Data travels a longer hazardous path from mobile device to the mobile cloud during offloading and thus consumes huge network bandwidth [2]. The bandwidth scarcity, and internet bottlenecks and traffic congestions are the catalysts for the higher latency of offloading computation. Real-time applications are highly latency sensitive and thus it requires to compute data in a close proximity of mobile devices or users. So, mobile fog can be the effective and suitable platform for offloading mobile computation.

Fog computing [3] is introduced by Cisco Systems Inc. to extend the cloud computing paradigm to the edge of network especially for Internet of Things (IoT) services. Mobile Fog is the complementary model of fog computing especially prototyped for seamless and latency-aware mobile services [4]. However, the key research questions for offloading computation in mobile fog are (1) How to offload computation in the mobile fog? (2) Which module or process of mobile application should offload? (3) Where to offload the module or process for minimizing the latency of service computing? Moreover, the mobility, heterogeneity and geographical distribution mobile devices impose additional challenges of computation offloading in mobile fog. This research contributes to finding the answer to the above questions. The key contributions of this research are as follows.

  • A code offloading framework is proposed for computation offloading in mobile fog environment. The code analyzer unit of the framework determines which basic blocks of the code are computation hungry and subject to offload.

  • A deep Q-learning [5] based computation offloading method is proposed for the autonomic management of massive offloading request. The trained code offloader unit of the proposed framework takes the offloading decision considering resource demand, resource availability and network status to minimize the latency of service computing.

  • The performance of the proposed model studied through simulation. The performance gain in terms of latency and energy efficiency justifies the dominance of the proposed autonomic offloading model.

Rest of the paper is organized as follows. In Section 2, we discussed the related works. The system model of mobile fog is presented in Section 3. The deep Q-learning based autonomic code offloading method in mobile fog is illustrated in Section 4. We presented the simulation and performance study results in Section 5. Finally, we concluded the paper in Section 6 with some future directions.

Section snippets

State-of-the-arts computation offloading methods

Mobile fog interplays with tradition cloud to access its huge computational resources. Thus, this section discussed state-of-the-arts resource provisioning methods of legacy cloud computing paradigm. Afterwards, the pioneer works on computation offloading in the mobile cloud are discussed in this section.

The elasticity and scalability of cloud computing are achieved through virtualization of cloud resources. The resources of cloud data centers are managed through VM configuration and placement

System model of mobile fog computing

Mobile Fog is the complementary model of fog computing especially prototyped for seamless and latency-aware mobile services. The system model of the mobile fog is presented in Fig. 1. The presented system model is derived from the hierarchical architecture of LTE (long-term evolution) 3GPP (3rd Generation Partnership Project) and Wi-Fi (wireless-fidelity) internetworking reference model [20].

In this architecture, the mobile fog is created on the edge of the networking modules. We consider the

Deep Q-learning based autonomic computation offloading

This section discusses the approach of offloading computation in mobile fog based on the system model presented in Section 3. The basic block [22] migration policy is used through mobile agents for offloading code from resource-constrained mobile stations to resource richer mobile fog. The programming code is partitioned into code generation and optimization unit of F-APC and deployed in different F-AP and also in different F-APC. According to the flow graph [22], [23], [24] of the generated

Performance evaluation

The performance of proposed computation offloading in mobile Fog is evaluated through simulation study. In the simulation topology, two adjacent mobile Fogs are connected and each of mobile fog contains one FNC or F-APC and three F-AP Fog nodes. Both of the Fog nodes are connected to the cloud node with eight VMs. We mostly focus on two performance criteria: response time and energy consumption. The simulation parameters are presented in Table 2. The flow graph of our studied benchmark

Conclusion

The proposed deep Q-learning based code offloading method leverage the mobile cloud computing. As it is a multi-agent based distributed method, agents learn from the environment through reinforcements. The offloading method deploys basic blocks in compatible fog nodes to support parallelism. The experimental results show the improved performance of the proposed offloading method in respect to execution time and latency and energy consumption.

Acknowledgment

The authors extend their appreciation to the Deanship of Scientific Research, King Saud University, Saudi Arabiafor funding this work through research group no (RGP- 1437-35).

Md Golam Rabiul Alam received B.S. and M.S. degrees in Computer Science and Engineering, and Information Technology respectively. He also received Ph.D. in Computer Engineering from Kyung Hee University, Korea in 2017. He is currently working as a Post-doctoral researcher in Computer Science and Engineering Department at Kyung Hee University, Korea. His research interest includes healthcare informatics, mobile cloud computing, ambient intelligence and persuasive technology.

References (24)

  • ZhengZ. et al.

    Coordinated Resource Provisioning and Maintenance Scheduling in Cloud Data Centers

    (2013)
  • BuX. et al.

    Coordinated self-configuration of virtual machines and appliances using a model-free learning approach

    IEEE Trans. Parallel Distrib. Syst.

    (2013)
  • Cited by (187)

    View all citing articles on Scopus

    Md Golam Rabiul Alam received B.S. and M.S. degrees in Computer Science and Engineering, and Information Technology respectively. He also received Ph.D. in Computer Engineering from Kyung Hee University, Korea in 2017. He is currently working as a Post-doctoral researcher in Computer Science and Engineering Department at Kyung Hee University, Korea. His research interest includes healthcare informatics, mobile cloud computing, ambient intelligence and persuasive technology.

    Mohammad Mehedi Hassan is currently an Associate Professor of Information Systems Department in the College of Computer and Information Sciences (CCIS), King Saud University (KSU), Riyadh, Kingdom of Saudi Arabia. He received his Ph.D. degree in Computer Engineering from Kyung Hee University, South Korea in February 2011. He received Best Paper Award from CloudComp conference at China in 2014. He also received Excellence in Research Award from CCIS, KSU in 2015 and 2016 respectively. He has published over 100+ research papers in the journals and conferences of international repute. He has served as, chair, and Technical Program Committee member in numerous international conferences/workshops like IEEE HPCC, ACM BodyNets, IEEE ICME, IEEE ScalCom, ACM Multimedia, ICA3PP, IEEE ICC, TPMC, IDCS, etc. He has also played role of the guest editor of several international ISI-indexed journals such as IEEE IoT, FGCS, etc. His research areas of interest are cloud federation, multimedia cloud, sensor-cloud, Internet of things, Big data, mobile cloud, cloud security, IPTV, sensor network, 5G network, social network, publish/subscribe system and recommender system. He is a member of IEEE.

    Md. Zia Uddin received his Ph.D. in Biomedical Engineering in February of 2011. He is currently working as a post-doctoral research fellow under Dept. of Informatics, University of Oslo, Norway. Dr. Zia’s researches are mainly focused on computer vision, image processing, artificial intelligence, and pattern recognition. He got more than 60 research publications including international journals, conferences and book chapters.

    Ahmad Almogren has received Ph.D. degree in computer sciences from Southern Methodist University, Dallas, Texas, USA in 2002. Previously, he worked as an assistant professor of computer science and a member of the scientific council at Riyadh College of Technology. He also served as the dean of the college of computer and information sciences and the head of the council of academic accreditation at Al Yamamah University. Presently, he works as an Associate Professor and the vice dean for the development and quality at the college of computer and information sciences at King Saud University in Saudi Arabia. He has served as a guest editor for several computer journals. His research areas of interest include mobile and pervasive computing, computer security, sensor and cognitive network, and data consistency.

    Giancarlo Fortino is currently a Professor of Computer Engineering (since 2006) at the Dept. of Informatics, Modeling, Electronics and Systems (DIMES) of the University of Calabria (Unical), Rende (CS), Italy. He holds the “Italian National Habilitation” for Full Professorship. He has been a visiting researcher at the International Computer Science Institute, Berkeley (CA), USA, in 1997 and 1999, and visiting professor at Queensland Univ. of Technology, Brisbane, Australia, in 2009.He was nominated Guest Professor in Computer Engineering of Wuhan Univ. of Technology (WUT) on April, 18 2012. His research interests include distributed computing, wireless sensor networks, software agents, cloud computing, Internet of Things systems. He authored over 230 publications in journals, conferences and books. He is the founding editor of the Springer Book Series on Internet of Things: Technology, Communications and Computing and serves in the editorial board of IEEE Transactions on Affective Computing, Journal of Networks and Computer Applications, Engineering Applications of Artificial Intelligence, Information Fusion, Multi Agent and GRID Systems, etc. He is co-founder and CEO of SenSysCal S.r.l., a spinoff of Unical, focused on innovative sensor-based systems for e-health and demotics. He is IEEE Senior member.

    View full text