隨著時代進步，裝置的種類及數量快速增長，裝置要處理的任務種類更是繁複多變，各個任務所需的執行時間長短不一，計算量可能超出裝置能負荷的能力，故無法滿足計算量大、且需要在時限內完成任務的需求。裝置可以利用雲端運算服務，將裝置上無法完成的任務，透過網路上傳至雲端伺服器，借助伺服器強大的運算能力，快速完成任務並將結果回傳。現今裝置接收任務的延遲需求更加的低，將任務傳輸到遙遠的雲端伺服器進行運算所需的額外傳輸延遲，可能會導致任務無法在時限內完成，因此有人提出了邊緣運算來解決這項問題。
邊緣運算為運算伺服器且更貼近裝置端，裝置透過更短的傳輸時間將任務傳送至伺服器，以達到任務延遲要求。以4G LTE為例，邊緣運算伺服器架設於4G基地台旁，使用者可以透過行動網路，將裝置的任務卸載(Offload)至邊緣運算伺服器做運算並將結果回傳，稱為行動邊緣運算(Mobile Edge Computing)。
本論文提出的Time Constraint-based Task Scheduling for Multi-Server (TCTSMS)，將邊緣運算伺服器接收的任務依據模糊理論，將任務分配至鄰近有空閒的伺服器進行運算，以降低自身負載，使邊緣運算伺服器接收並完成更多的任務、降低任務拒絕率。任務接收策略分為兩種方法：搶奪(Preemptive)以及非搶奪(Non-preemptive)，搶奪可讓延遲需求較緊急的任務插隊於能容忍較長延遲的任務之前，使較緊急的任務能及早執行，不會因為被能容忍較長延遲的任務較多，而造成排隊時間過長或因此被伺服器拒絕接受任務；本論文假設了任務提出的計算需求與實際執行時會有所差異而可能導致超時，更能貼近現實任務卸載的不確定因素。

As the evolution of technology, more and more devices are innovated. Device have wide variety of task. Some Tasks need to be executed in delay constraint time. Device may not satisfy task’s demand of delay because of low CPU capacity of device. If device can not satisfy task’s demand, device can offload the task to cloud server through network. Offload task to cloud will cause transmission time between server and device. Cloud server has powerful CPU to help device executed task, but the transmission time is too long to fulfill delay constraint.
Edge Computing is a kind of offloading method but more closed to user. For instance, Mobile Edge computing will deploy the server near by the LTE base station. Device can offload task to edge server through 4G LTE network. Task only needs short transmission time to complete delay sensitive task by offload task to edge server.
This paper proposes Time Constraint-based Task Scheduling for Multi-Server (TCTSMS) algorithm. TCTSMS Algorithm use fuzzy theory to choose task will be executed at local server or offload to other idle server with set of edge server. If one server has many offload task requests as hotspot, total throughput will increase and blocking rate decrease. This paper proposes preemptive method to ensure delay sensitive task would not be blocked because of lots of task with long delay tolerance occupy edge computing system. This paper assumes that the task provides information of required execution cycle is not equals to real execution cycle. This uncertainty situation may be closer to reality.

[1] 3GPP, TS 36.300 V10.4.0, Evolved Universal Terrestrial Radio Access.
[2] 許亨仰, Communication Components Magazine, [Online]. Available: https://www.2cm.com.tw/2cm/zh-tw/tech/848B4DFB7B584537B42D0006E1EB6523. [Accessed 29 5 2021].
[3] [Online]. Available: https://en.wikipedia.org/wiki/System_Architecture_Evolution. [Accessed 29 05 2021].
[4] Keysight Technologies, Inc., [Online]. Available: http://rfmw.em.keysight.com/wireless/helpfiles/89600b/webhelp/subsystems/lte/content/lte_overview.htm. [Accessed 30 05 2021].
[5] P. Panigrahi, "3GLTEinfo," 3GLTEInfo, 6 12 2013. [Online]. Available: https://www.3glteinfo.com/intra-lte-handover-using-s1-interface/. [Accessed 30 5 2021].
[6] P. Panigrahi, "3GLTEinfo," 3GLTEinfo, 4 12 2013. [Online]. Available: https://www.3glteinfo.com/intra-lte-handover-using-x2-interface/. [Accessed 30 5 2021].
[7] Z. Liang, Y. Liu, T. M. Lok and K. Huang, "Multi-Cell Mobile Edge Computing: Joint Service Migration and Resource Allocation," IEEE Transactions on Wireless Communications, p. DOI: 10.1109/TWC.2021.3070974, 12 Apr. 2021.
[8] Y. Qiu, H. Zhang, K. Long, H. Sun, X. Li and V. C. Leung, "Improving handover of 5G networks by network function virtualization and fog computing," in IEEE/CIC International Conference on Communications in China (ICCC), Qingdao, China, 2017.
[9] Microsoft, "Describe different cloud services," 2021. [Online]. Available: https://docs.microsoft.com/EN-US/learn/modules/fundamental-azure-concepts/. [Accessed 31 5 2021].
[10] K. Dolui and S. K. Datta, "Comparison of edge computing implementations: Fog computing, cloudlet and mobile edge computing," in Global Internet of Things Summit (GIoTS), Geneva, Switzerland, 2017.
[11] I. Elgendy, W. Zhang, Y.-C. Tian and K. Li, "Resource allocation and computation offloading with data security for mobile edge computing," Future Generation Computer Systems, pp. 531-541, 2019.
[12] Z. Pang, L. Sun, Z. Wang, E. Tian and S. Yang, "A Survey of Cloudlet based Mobile Computing," in International Conference on Cloud Computing and Big Data, Shanghai, China, 2015.
[13] J.Saton, Wauters, T. Volckaert and B. T. F. D., "Fog Computing: Enabling the Management and Orchestration of Smart City Applications in 5G Networks," Entropy, doi:10.3390/e20010004, 20 Apr. 2018.
[14] L. A. Zadeh, "Fuzzy Sets," Information and Control, pp. 338-353, 1965.
[15] 楊敏生, "模糊理論簡介," in 數學傳播, 中央研究院數學研究所, 1994, pp. 1-5.
[16] J. V. Joseph, J. Kwak and G. Iosifidis, "Dynamic Computation Offloading in Mobile-Edge-Cloud Computing Systems," in IEEE Wireless Communications and Networking Conference, Marrakesh, Morocco, 2019.
[17] J. Meena, M. Kumar and M. Vardhan, "Cost Efficient Scheduling for Delay-sensitive Tasks in Edge Computing System," in IEEE International Conference on Services Computing, San Francisco, CA, USA, 2018.
[18] B. Dab, N. Aitsaadi and R. Langar, "Joint Optimization of Offloading and Resource Allocation Scheme for Mobile Edge Computing," in IEEE Wireless Communications and Networking Conference, Marrakesh, Morocco, 2019.
[19] K. Guo and T. Q. S. Quek, "Dynamic Computation Offloading in Multi-Server MEC Systems: An Online Learning Approach," in GLOBECOM 2020 - 2020 IEEE Global Communications Conference, Taipei, Taiwan, 2020.
[20] D. Wang, X. Tian, H. Cui and Z. Liu, "Reinforcement learning-based joint task offloading and migration schemes optimization in mobility-aware MEC network," China Communications, pp. 31-44, 9 Sep. 2020.
[21] C. Liu, F. Tang, Y. Hu, K. Li, Z. Tang and L. K., "Distributed Task Migration Optimization in MEC by Extending Multi-Agent Deep Reinforcement Learning Approach," IEEE Transactions on Parallel and Distributed Systems, vol. 32, no. 7, 1 Jul. 2021.
[22] W. C. Chang, Y. L. Chen and P. C. Wang, "Hotspot Mitigation for Mobile Edge Computing," IEEE Transactions on Sustainable Computing, doi:10.1109/TSUSC.2018.2878438, 29 Oct. 2018.