This paper studies a problem of jointly optimizing two important operations in mobile edge computing: service caching, which determines which services to be hosted at the edge, and service routing, which determines which requests to be processed at the edge. We aim to address several practical challenges, including limited storage and computation capacities of edge servers, delay of reconfiguring edge servers, and unknown future request arrival patterns. To this end, we formulate the problem as an online optimization problem, in which the objective function includes both the costs of forwarding requests, processing requests, and reconfiguring edge servers. By leveraging a natural timescale separation between service routing and service caching, namely, the former happens faster than the latter, we propose an online two-stage algorithm and its randomized variant. Both algorithms have low complexity and our fractional solution achieves sublinear regret. Simulation results show that our algorithms significantly outperform other state-of-the-art policies, including one that assumes the knowledge of all future request arrivals.

翻译：本文研究的是共同优化移动边缘计算中两个重要操作的问题:服务缓存,它决定了边缘的托管服务,服务路由,它决定了边缘的处理请求。我们的目标是应对若干实际挑战,包括边缘服务器的储存和计算能力有限、边缘服务器的重新配置延迟以及未知的未来请求抵达模式。为此,我们将这一问题描述为一个在线优化问题,其中目标功能包括传输请求、处理请求和重新配置边缘服务器的费用。通过利用服务路线和服务缓存之间的自然时间尺度分离,即前者发生速度快于后者,我们提议采用在线双阶段算法及其随机变式。两种算法的复杂程度都较低,我们的零碎解决方案都实现了亚线性遗憾。模拟结果显示,我们的算法大大超过其他最先进的政策,包括假设未来所有请求抵达者知识的算法。