Modeling and Analyzing Offloading Strategies of IoT Applications over Edge Computing and Joint Clouds
Abstract
:1. Introduction
- Propose a performance-driven method to measure the end-to-end IoT service effectiveness, taking both computational and communication resources into account;
- Conduct an asymptotic performance analysis to deeply dive into system behavior under the loosely-coupled and the orchestrator-enabled offloading schemes;
- Provide a set of models and algorithms in order to handle various IoT applications’ requirements and achieve an optimal offloading strategy;
- An evaluation of the IoT task offloading strategies under different constraints, which can be employed in improving the efficiency of task offloading and achieving well-balanced resource management in Edge-Cloud environments.
2. Related Work
3. Background and Problem Motivation
3.1. Resource and Workloads in IoT Scenario
3.1.1. Resource Entities in Edge and Cloud Environments
3.1.2. IoT Workloads
3.1.3. IoT Deployment and Task Offloading
3.2. Research Problem and Challenges
3.2.1. Problem Statement
3.2.2. Emerging Challenges
- Scale and Complexity: with the increase of IoT manufacturers developing heterogeneous sensors and smart devices, selecting optimal resources to hold IoT tasks over the joint cloud environment becomes increasingly complicated when considering customized hardware configurations and personalized requirements. For example, some tasks can only operate with specific hardware architectures (e.g., ARM and Intel) or operating systems, while tasks with high-security requirements might require specific hardware and protocols to function. Not only does orchestration cater to such functional requirements, but it must also do so in the face of increasingly larger workflows that change dynamically. The orchestrator must determine whether the assembled systems comprising of Cloud resources, Edge nodes and end-user devices coupled with geographic distributions and constraints are capable of provisioning complex services correctly and efficiently. In particular, the orchestrator must be able to automatically predict, detect and resolve scalability bottlenecks which may arise from an increased application scale.
- Dynamicity: this is one of the main characteristics of IoT applications, whose topology or diverse resources may change dynamically. This is a particular problem in the context of software upgrades or frequent join-leave behavior of network objects which will change its internal properties and performance, potentially altering the overall workload pattern. Specifically, the communication link is likely to be influenced by the fluctuation of bandwidth, connectivity and device mobility. It may lead to unpredictable demands of task offloading and resource allocation over the joint Clouds and Edge nodes.
- Scalability: scalability is another aspect of the challenge. In this context, it specifically refers to the capability of handling an explosive number of IoT applications and dynamic resource changes. Additionally, attributes of IoT tasks dynamically change, thus each task’s procedure may have different execution times. In addition, IoT devices are mobile, the number of devices may increase in some areas, thus the workload will be increased for the connected edge node. Hence, the amount of IoT workload will change dynamically over Edge-Cloud system, which could lead to service performance degradation.
4. Methodology
4.1. Mainstream Architectural Schemes
4.2. Offloading Strategies
4.2.1. Static-Ratio Policy
Algorithm 1 Static Ratio Policy |
Require: : the latest ratio of tasks that have been offloaded into the current edge node. |
Ensure: or |
1: if and then |
2: |
3: |
⊳dispatch into Edge node |
4: return |
5: else |
6: |
⊳offload into Cloud |
7: return |
8: end if |
4.2.2. Least-Load Policy
Algorithm 2 Least Load Policy |
Require: : total Edge-Cloud machines with load metric |
Ensure: : optimal machine with least load |
1: while .location is null do |
2: |
3: if then |
4: |
5: |
⊳dispatch into machine |
6: return |
7: else |
8: |
9: end if |
10: end while |
4.2.3. Probabilistic Policy
Algorithm 3 Probabilistic Policy | |
Require:: a matrix of conditional probabilities. | |
Ensure:: optimal machine with least cost including utilization, transmission, etc. for | |
1: | |
⊳outputs a matrix consisting of conditional probabilities of every pair of tasks, depending on counting occurrence of and | |
2: | |
3: for all do | |
4: | |
5: end for | |
6: | |
7: multilevel sorting () | |
⊳sort as delay-sensitivity, edge-util, cloud-util, input-size... | |
8: for all do | |
9: let null | |
10: if t is delay-sensitive then | |
11: | |
⊳an edge node with the least cost for task t | |
12: else | |
13: | |
⊳a cloud node with the least cost for task t | |
14: end if | |
15: if t is then | |
16: | ⊳dispatch t into |
17: return | |
18: else | |
19: virtually delete allocated resources | |
20: end if | |
21: end for |
4.3. Measurement of Performance
4.3.1. End-to-End Service Time
4.3.2. Measurement
- Latency to local Edge: the total service time for offloading tasks to connected Edge node in Edge-Cloud system is the sum of: time to send task data to the Edge server, queuing to process the task in the Edge server, the processing time of the task and the time to send the output to the IoT devices. We consider the connection to the local Edge will be within WLAN, as presented in Equation (1).
- Latency to collaborative Edge: the total service time to process an offloading task in a collaborative Edge can be calculated by the summation of time of uploading task data to local Edge via WLAN, uploading data to the collaborative Edge via MAN, queuing time, processing time and time for downloading the output to IoT device via MAN and WLAN, as presented in Equation (2).
- Latency to central Cloud: in order to calculate service time for an offloading task in the cloud, we should consider the network delay through WLAN, MAN, WAN as well as the queuing time and processing time, as presented in Equation (3).
5. Evaluation
5.1. Simulation Set-Up
5.2. Methodology
5.3. Service Time vs. Resource Requirements
5.4. Scalability and Availability
6. Research Limitations
- This work handles the offloading strategies of independent tasks; however, task dependency plays an essential factor to affect the decision of offloading tasks. Thus, this work can be extended to consider task dependency in the process of scheduling offloading tasks. Task dependency and the intercommunication between tasks can be represented as a DAG, which can be modeled within the proposed approach to enhance the overall service time of latency-sensitive applications.
- Another complement work that will enhance the work presented is to predict the behaviour of latency-sensitive applications [25,26]. The prediction can be in several areas such as predicting the volume of incoming tasks, predicting the users’ mobility which could help to determine their locations. Therefore, it would help the resource manager to prepare the required resources in advance and avoided any performance degradation. This extension would be useful when scheduling offloading tasks in order to minimise the overall service time.
- In addition, the proposed approach considers three customized offloading strategies in order to handle various requirements for IoT latency-sensitive applications based on CPU speed and bandwidth. Thus, the approach could be extended to consider computational resources such as different GPUs and FPGAs since there are many applications for AR/VR and video gaming that require intensive computations in order to process their tasks in the Edge-Cloud environment.
7. Conclusions and Future Work
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- IoT Number Trend. Available online: https://www.businessinsider.com/internet-of-everything-2015-bi-2014-12 (accessed on 13 August 2020).
- Yousefpour, A.; Fung, C.; Nguyen, T.; Kadiyala, K.; Jalali, F.; Niakanlahiji, A.; Kong, J.; Jue, J.P. All one needs to know about fog computing and related edge computing paradigms: A complete survey. J. Syst. Archit. 2019, 98, 289–330. [Google Scholar] [CrossRef]
- Vermesan, O.; Friess, P.; Guillemin, P.; Gusmeroli, S.; Sundmaeker, H.; Bassi, A.; Jubert, I.S.; Mazura, M.; Harrison, M.; Eisenhauer, M.; et al. Internet of things strategic research roadmap. Internet Things Glob. Technol. Soc. Trends 2011, 1, 9–52. [Google Scholar]
- Gubbi, J.; Buyya, R.; Marusic, S.; Palaniswami, M. Internet of Things (IoT): A vision, architectural elements, and future directions. Future Gener. Comput. Syst. 2013, 29, 1645–1660. [Google Scholar] [CrossRef] [Green Version]
- Yi, S.; Li, C.; Li, Q. A survey of fog computing: Concepts, applications and issues. In Proceedings of the 2015 Workshop on Mobile Big Data; ACM: New York, NY, USA, 2015; pp. 37–42. [Google Scholar]
- Shekhar, S.; Gokhale, A. Dynamic resource management across cloud-edge resources for performance-sensitive applications. In Proceedings of the 17th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing, Madrid, Spain, 14–17 May 2017; pp. 707–710. [Google Scholar]
- Satyanarayanan, M.; Bahl, V.; Caceres, R.; Davies, N. The case for vm-based cloudlets in mobile computing. IEEE Pervas. Comput. 2009, 8, 14–23. [Google Scholar] [CrossRef]
- Zhang, Y.; Yang, R.; Wo, T.; Hu, C.; Kang, J.; Cui, L. Cloudap: Improving the qos of mobile applications with efficient vm migration. In Proceedings of the 2013 IEEE 10th International Conference on High Performance Computing and Communications & 2013 IEEE International Conference on Embedded and Ubiquitous Computing, Zhangjiajie, China, 13–15 November 2013; pp. 1374–1381. [Google Scholar]
- Toczé, K.; Nadjm-Tehrani, S. A taxonomy for management and optimization of multiple resources in edge computing. Wirel. Commun. Mob. Comput. 2018, 2018, 7476201. [Google Scholar] [CrossRef] [Green Version]
- Premsankar, G.; Di Francesco, M.; Taleb, T. Edge computing for the Internet of Things: A case study. IEEE Internet Things J. 2018, 5, 1275–1284. [Google Scholar] [CrossRef] [Green Version]
- Li, W.; Santos, I.; Delicato, F.C.; Pires, P.F.; Pirmez, L.; Wei, W.; Song, H.; Zomaya, A.; Khan, S. System modelling and performance evaluation of a three-tier Cloud of Things. Future Gener. Comput. Syst. 2017, 70, 104–125. [Google Scholar] [CrossRef]
- Hegyi, A.; Flinck, H.; Ketyko, I.; Kuure, P.; Nemes, C.; Pinter, L. Application orchestration in mobile edge cloud: Placing of iot applications to the edge. In Proceedings of the 2016 IEEE 1st International Workshops on Foundations and Applications of Self Systems, Augsburg, Germany, 12–16 September 2016; pp. 230–235. [Google Scholar]
- Dinh, T.Q.; Tang, J.; La, Q.D.; Quek, T.Q. Offloading in mobile edge computing: Task allocation and computational frequency scaling. IEEE Trans. Commun. 2017, 65, 3571–3584. [Google Scholar]
- Du, J.; Zhao, L.; Feng, J.; Chu, X. Computation offloading and resource allocation in mixed fog/cloud computing systems with min-max fairness guarantee. IEEE Trans. Commun. 2018, 66, 1594–1608. [Google Scholar] [CrossRef] [Green Version]
- Liu, J.; Mao, Y.; Zhang, J.; Letaief, K.B. Delay-optimal computation task scheduling for mobile-edge computing systems. In Proceedings of the 2016 IEEE International Symposium on Information Theory (ISIT), Barcelona, Spain, 10–15 July 2016; pp. 1451–1455. [Google Scholar]
- Rodrigues, T.G.; Suto, K.; Nishiyama, H.; Kato, N. Hybrid method for minimizing service delay in edge cloud computing through VM migration and transmission power control. IEEE Trans. Comput. 2016, 66, 810–819. [Google Scholar] [CrossRef]
- Wei, X.; Tang, C.; Fan, J.; Subramaniam, S. Joint Optimization of Energy Consumption and Delay in Cloud-to-Thing Continuum. IEEE Internet Things J. 2019, 6, 2325–2337. [Google Scholar] [CrossRef]
- Yang, B.; Chai, W.K.; Pavlou, G.; Katsaros, K.V. Seamless support of low latency mobile applications with nfv-enabled mobile edge-cloud. In Proceedings of the 2016 5th IEEE International Conference on Cloud Networking (Cloudnet), Pisa, Italy, 3–6 October 2016; pp. 136–141. [Google Scholar]
- Zeng, D.; Gu, L.; Guo, S.; Cheng, Z.; Yu, S. Joint optimization of task scheduling and image placement in fog computing supported software-defined embedded system. IEEE Trans. Comput. 2016, 65, 3702–3712. [Google Scholar] [CrossRef]
- Wang, S.; Zafer, M.; Leung, K.K. Online placement of multi-component applications in edge computing environments. IEEE Access 2017, 5, 2514–2533. [Google Scholar] [CrossRef]
- Tärneberg, W.; Mehta, A.; Wadbro, E.; Tordsson, J.; Eker, J.; Kihl, M.; Elmroth, E. Dynamic application placement in the mobile cloud network. Future Gener. Comput. Syst. 2017, 70, 163–177. [Google Scholar] [CrossRef]
- Fan, Q.; Ansari, N. Application aware workload allocation for edge computing-based IoT. IEEE Internet Things J. 2018, 5, 2146–2153. [Google Scholar] [CrossRef]
- Roy, D.G.; De, D.; Mukherjee, A.; Buyya, R. Application-aware cloudlet selection for computation offloading in multi-cloudlet environment. J. Supercomput. 2017, 73, 1672–1690. [Google Scholar] [CrossRef]
- Mahmud, R.; Ramamohanarao, K.; Buyya, R. Application Management in Fog Computing Environments: A Taxonomy, Review and Future Directions. ACM Comput. Surv. 2020, 53, 1–43. [Google Scholar] [CrossRef]
- Ijaz, M.F.; Attique, M.; Son, Y. Data-Driven Cervical Cancer Prediction Model with Outlier Detection and Over-Sampling Methods. Sensors 2020, 20, 2809. [Google Scholar] [CrossRef]
- Ali, F.; El-Sappagh, S.; Islam, S.R.; Kwak, D.; Ali, A.; Imran, M.; Kwak, K.S. A smart healthcare monitoring system for heart disease prediction based on ensemble deep learning and feature fusion. Inf. Fusion 2020, 63, 208–222. [Google Scholar] [CrossRef]
- Wen, Z.; Yang, R.; Garraghan, P.; Lin, T.; Xu, J.; Rovatsos, M. Fog orchestration for internet of things services. IEEE Internet Comput. 2017, 21, 16–24. [Google Scholar] [CrossRef] [Green Version]
- Sharma, S.K.; Wang, X. Live data analytics with collaborative edge and cloud processing in wireless IoT networks. IEEE Access 2017, 5, 4621–4635. [Google Scholar] [CrossRef]
- Yousefpour, A.; Ishigaki, G.; Jue, J.P. Fog computing: Towards minimizing delay in the internet of things. In Proceedings of the 2017 IEEE international conference on edge computing (EDGE), Honolulu, HI, USA, 25–30 June 2017; pp. 17–24. [Google Scholar]
- Wang, N.; Varghese, B.; Matthaiou, M.; Nikolopoulos, D.S. ENORM: A framework for edge node resource management. IEEE Trans. Serv. Comput 2017, 13, 1086–1099. [Google Scholar] [CrossRef] [Green Version]
- Wang, J.; Pan, J.; Esposito, F.; Calyam, P.; Yang, Z.; Mohapatra, P. Edge cloud offloading algorithms: Issues, methods, and perspectives. ACM Comput. Surv. (CSUR) 2019, 52, 2. [Google Scholar] [CrossRef]
- Farris, I.; Taleb, T.; Flinck, H.; Iera, A. Providing ultra-short latency to user-centric 5G applications at the mobile network edge. Trans. Emerg. Telecommun. Technol. 2018, 29, e3169. [Google Scholar] [CrossRef] [Green Version]
- Li, C.; Tang, J.; Tang, H.; Luo, Y. Collaborative cache allocation and task scheduling for data-intensive applications in edge computing environment. Future Gener. Comput. Syst. 2019, 95, 249–264. [Google Scholar] [CrossRef]
- Mahmud, R.; Koch, F.L.; Buyya, R. Cloud-fog interoperability in IoT-enabled healthcare solutions. In Proceedings of the 19th International Conference on Distributed Computing and networking, Varanasi, India, 4–7 January 2018; p. 32. [Google Scholar]
- Sonmez, C.; Ozgovde, A.; Ersoy, C. Edgecloudsim: An environment for performance evaluation of edge computing systems. Trans. Emerg. Telecommun. Technol. 2018, 29, e3493. [Google Scholar] [CrossRef]
- Calheiros, R.N.; Ranjan, R.; Beloglazov, A.; De Rose, C.A.; Buyya, R. CloudSim: A toolkit for modeling and simulation of cloud computing environments and evaluation of resource provisioning algorithms. Software Pract. Exper 2011, 41, 23–50. [Google Scholar] [CrossRef]
- Sun, X.; Hu, C.; Yang, R.; Garraghan, P.; Wo, T.; Xu, J.; Zhu, J.; Li, C. ROSE: Cluster resource scheduling via speculative over-subscription. In Proceedings of the 2018 IEEE 38th International Conference on Distributed Computing Systems (ICDCS), Vienna, Austria, 2–5 July 2018; pp. 949–960. [Google Scholar]
Parameter | Value |
---|---|
Poisson Interarrival Time of Tasks (second) | 3 |
Simulation time (hours) | 2 |
Warm up period (s) | 3 |
Number of repetitions | 5 |
Number of Edge Nodes | 3 |
Number of hosts per edge nodes | 2 |
Number of VMs per Edge server/cloud | 4/ not limited |
VM speed (MIPS) per Edge server/cloud | 10,000 |
Minimum number of end devices | 100 |
Maximum number of end devices | 1000 |
Active/idle for end devices period (s) | 45/15 |
Probability of Offloading to Cloud (%) | 10 |
Average Data Size for Upload/Download (KB) | 1500/15 |
Average Task Size (MI) | 1500 |
WAN/WLAN Bandwidth (Mbps) | 20/300 |
WAN Propagation Delay (ms) | 100 |
500 MIPS | 1000 MIPS | 2000 MIPS | 4000 MIPS | |
---|---|---|---|---|
0.25 MB | App1 | App2 | App3 | App4 |
0.5 MB | App5 | App6 | App7 | App8 |
0.75 MB | App9 | App10 | App11 | App12 |
1 MB | App13 | App14 | App15 | App16 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Share and Cite
Almutairi, J.; Aldossary, M. Modeling and Analyzing Offloading Strategies of IoT Applications over Edge Computing and Joint Clouds. Symmetry 2021, 13, 402. https://doi.org/10.3390/sym13030402
Almutairi J, Aldossary M. Modeling and Analyzing Offloading Strategies of IoT Applications over Edge Computing and Joint Clouds. Symmetry. 2021; 13(3):402. https://doi.org/10.3390/sym13030402
Chicago/Turabian StyleAlmutairi, Jaber, and Mohammad Aldossary. 2021. "Modeling and Analyzing Offloading Strategies of IoT Applications over Edge Computing and Joint Clouds" Symmetry 13, no. 3: 402. https://doi.org/10.3390/sym13030402