T2TRG Hong, Choong Seon Internet-Draft Kyung Hee University Intended status: Standards Track Chit Wutyee Zaw Expires: August 09, 2020 Kyung Hee University Kang, Seok Won Kyung Hee University October 2020 User Centric Assignment and Partial Task Offloading for Mobile Edge Computing in Ultra-Dense Networks draft-hongcs-chit-mec-00 Abstract By collocating servers at base stations, Mobile Edge Computing (MEC) provides low latency to users for real time applications such as Virtual Reality and Augmented Reality. To satisfy the growing demand of users, base stations are deployed densely in highly populated areas. Coordinated Multipoint Transmission (CoMP) allows users to connect to multiple base stations simultaneously. In ultra-dense networks, by offloading the partials of tasks to different base stations, users can achieve lower latency and utilize the computation ability of the surrounding base stations. To control the signaling overhead, the number of base stations that can be connected should be limited. In this paper, we propose a user-centric base station assignment algorithm by considering the possible load of base stations. Moreover, a partial task offloading algorithm is proposed to utilize the computation of under-loaded base stations. Resource allocation is then solved by convex optimization. Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." Hong, et al. Expires August 09, 2020 [Page 1] Internet-Draft Task Offloading for MEC October 2020 This Internet-Draft will expire on August 09, 2020. Copyright Notice Copyright (c) 2018 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License. Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1. Terminology and Requirements Language . . . . . . . . . . 2 2. System Model . . . . . . . . . . . . . . . . . . . . . . . . . . 2 3. Problem Formulation. . . . . . . . . . . . . . . . . . . . . . . 4 4. User-centric Assignment and Partial Offloading . . . . . . . . . 3 4.1. User-centric Assignment. . . . . . . . . . . . . . . . . 3 4.2. Partial Offloading . . . . . . . . . . . . . . . . . . . . . . 4 4.3. Radio Resource Allocation. . . . . . . . . . . . . . . . . . . 5 5. Results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 6. IANA Considerations. . . . . . . . . . . . . . . . . . . . . . . 6 7. Security Considerations . . . . . . . . . . . . . . . . . . . . 6 8. References . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 8.1. Normative References . . . . . . . . . . . . . . . . . . . . . 6 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . . 7 1. Introduction Mobile Edge Computing (MEC) has been an interesting topic in both academia and industry for its ability to provide low latency and high computation to users by setting up severs near to users. Computation and latency intensive applications requires users to offload their tasks to servers to achieve the minimum delay and maintain the energy of users’ devices. In densely deployed networks, users can utilize the resources of nearby base stations (BS) by offloading partials of their tasks with the technology provided by Coordinated Multipoint Transmission (CoMP). Despite the advantages that MEC brings, there are many challenges to tackle in MEC which are pointed out in [1]. The communication aspect is surveyed in [2] where authors considered joint management of radio and computation resources. Authors also introduced standards and application scenarios. Hong, et al. Expires August 09, 2020 [Page 2] Internet-Draft Task Offloading for MEC October 2020 Authors in [3] developed a distributed approach for the offloading of computation tasks, caching of content and allocation of resources by using an alternating direction method of multipliers. Task offloading for ultra-dense network was considered in [4] where authors divided the task placement and resource allocation problems and proposed an efficient offloading approach. But, authors considered to offload to one BS. In this paper, we consider partial offloading in ultra-dense networks. To avoid the overloading at BSs, we take the number of possible users who can connect to BSs into account and propose a heuristic algorithm for user-centric assignment. In addition, a partial offloading algorithm is proposed to utilize the resources of under- loaded BSs by offloading the larger portion of tasks to those BSs. Then, resource allocation is solved with the help of convex optimization. 1.1. Terminology and Requirements Language The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119]. 2. System Model A network with densely deployed BSs is considered where users can offload their tasks to multiple BSs simultaneously. We consider the Orthogonal Frequency Division Multiple Access in both uplink and downlink transmission. We also consider that MEC server are equipped with multi-core technology that they can compute offloaded tasks simultaneously. The user’s task has three parameters such as the size of input file, output result and task in CPU cycles. Hong, et al. Expires August 09, 2020 [Page 3] Internet-Draft Task Offloading for MEC October 2020 +---------------+ +--------+ +--------+ +--------+ | Mobile Device | | SBS 1 | | SBS 2 | | SBS 3 | +---------------+ +--------+ +--------+ +--------+ | | | +-----------------+ | | | Offload partial | | | | portion of task | | | +-----------------+ | | +-----------------+ | | Compute the | | | offloaded task | | +-----------------+ | +-----------------+ | | | Return task | | | | result | | | +-----------------+ | --------------------------------------------------------------- | +-----------------+ | | | Offload partial | | | | portion of task | | | +-----------------+ | | +-----------------+ | | Compute the | | | offloaded task | | +-----------------+ | +-----------------+ | | | Return task | | | | result | | | +-----------------+ | --------------------------------------------------------------- | +-----------------+ | | | Offload partial | | | | portion of task | | | +-----------------+ | | +-----------------+ | | Compute the | | | offloaded task | | +-----------------+ | +-----------------+ | | | Return task | | | | result | | | +-----------------+ | --------------------------------------------------------------- Figure 1: Partial offloading with Coordinated Transmission in an Ultra-Dense Network Hong, et al. Expires August 09, 2020 [Page 4] Internet-Draft Task Offloading for MEC October 2020 3. Problem Formulation The objective of the partial offloading and resource allocation problem is to minimize the latency of all mobile users where the task must be computed fully. The maximum number of SBSs that a user can associate to is limited. The uplink bandwidth for task offloading and downlink bandwidth for result transmission are limited. In addition, the computing resource at MEC servers and local computing resource are also restricted. 4. User-centric Assignment and Partial Offloading 4.1. User-centric Assignment to SBSs First, we need to determine the user assignment to the BSs by considering the overloading possibility. The score from a user to a SBS is calculated in which the uplink, downlink singal-to-noise ratios and the inverse proportion of the number of users who are likely to associate to a SBS is considered. +---------------+ +--------+ +--------+ +--------+ | Mobile Device | | SBS 1 | | SBS 2 | | SBS 3 | +---------------+ +--------+ +--------+ +--------+ | +-----------------+ | Calculate score | | for all SBSs | +-----------------+ | +-----------------+ | Choose 3 SBSs | | with highest | | scores | +-----------------+ | --------------------------------------------------------------- | +-----------------+ | | | Send the signal | | | | for assignment | | | +-----------------+ | --------------------------------------------------------------- | +-----------------+ | | | Send the signal | | | | for assignment | | | +-----------------+ | --------------------------------------------------------------- | +-----------------+ | | | Send the signal | | | | for assignment | | | +-----------------+ | --------------------------------------------------------------- Figure 2: User-centric Assignment Hong, et al. Expires August 09, 2020 [Page 5] Internet-Draft Task Offloading for MEC October 2020 4.2 Partial Task Offloading After the assignment is done, the fractions of the task allocated to BSs are resolved by utilizing the resources of under-loaded BSs. The higher portion of a task is offloaded to a SBS with a lower total computing load of all the assigned users. SBSs are sorted according to the increasing computing loads of the users. The portion of the task is offloaded to SBSs in the order. +------------------+---------------+------------+------------+ User's| Portion | Portion | Portion | Portion | task | offloaded to | offloaded to |offloaded to| computed at| | SBS 1 | SBS 2 | SBS 3 | the user | +------------------+---------------+------------+------------+ \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \ / \/ \/ +------------+ +-----------+ +-----------+ +-----------+ | SBS 1 | | SBS 2 | | SBS 3 | | Local | +------------+ +-----------+ +-----------+ +-----------+ Figure 3: Partial Task Offloading 4.3. Radio Resource Allocation After obtaining the partial task offloading, we need to solve the resource allocation problem. The resource allocation problem is convex which can easily be solved. In this paper, we use cvxpy [5] to solve this problem. For the local CPU cycles assignment, the maximum available CPU cycle is assigned since the objective is minimizing the latency. Hong, et al. Expires August 09, 2020 [Page 6] Internet-Draft Task Offloading for MEC October 2020 5. Results Poisson Point Process is used to model the deployment of BSs and users where their densities are 0.6/m2 6/m2 respectively. For power density thermal noise, -174dBm/Hz is used. Fig. 2 shows the simulation setup used in the paper. Transmit power of pico BSs and users are 23dbm and 20dbm respectively. CPU speed is 4GHz at MEC server and 0.3GHz at user. The total uplink and downlink bandwidth are 20MHz each. The size of input file follows a uniform distribution between 300 and 800 KB. The uniform distribution is also used to model the size of tasks and output files which are 0.5 to 1 GHz and 0.2 to 2.5 MB respectively. The latency obtained at SBSs are different but most of the SBSs have the similar latency results due to the different user task requirements. In the highly dense networks, the proposed approach can keep most of the BSs to achieve comparable results. The proposed approach obtains lower latency compared to the baseline approach where the loads of SBSs are not considered and task allocation is done uniformly. The difference becomes significant as the number of users increases. 6. IANA Considerations There are no IANA considerations related to this document. 7. Security Considerations There are no security considerations related to this document. 8. References 8.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [1] P. Mach and Z. Becvar, "Mobile Edge Computing: A Survey on Architecture and Computation Offloading," IEEE Communications Surveys & Tutorials, vol. 19, no. 3, pp. 1628-1656, 2017. [2] Y. Mao, C. You, J. Zhang, K. Huang and K. B. Letaief, "A Survey on Mobile Edge Computing: The Communication Perspective," IEEE Communications Surveys & Tutorials, vol. 19, no. 4, pp. 2322-2358, 2017. [3] C. Wang, C. Liang, F. R. Yu, Q. Chen and L. Tang, "Computation Offloading and Resource Allocation in Wireless Cellular Networks With Mobile Edge Computing," IEEE Transactions on Wireless Communications, vol. 16, no. 8, pp. 4924-4938, 2017. [4] M. Chen and Y. Hao, "Task Offloading for Mobile Edge Computing in Software Defined Ultra-Dense Network," IEEE Journal on Selected Areas in Communications, vol. 36, no. 3, pp. 587-597, 2018. [5] S. Diamond and S. Boyd, "CVXPY: A Python-Embedded Modeling Language for Convex Optimization," Journal of Machine Learning Research, vol. 17, no. 83, pp. 1-5, 2016. Hong, et al. Expires August 09, 2020 [Page 7] Internet-Draft Task Offloading for MEC October 2020 Authors' Addresses Choong Seon Hong Computer Science and Engineering Department, Kyung Hee University Yongin, South Korea Phone: +82 (0)31 201 2532 Email: cshong@khu.ac.kr Chit Wutyee Zaw Computer Science and Engineering Department, Kyung Hee University Yongin, South Korea Phone: +82 (0)31 201 2987 Email: cwyzaw@khu.ac.kr Seok Won Kang Computer Science and Engineering Department, Kyung Hee University Yongin, South Korea Phone: +82 (0)31 201 2987 Email: dudtntdud@khu.ac.kr