Virtual Servers Co-Migration for Mobile Accesses: Online versus Off-Line

Abstract

In this paper, we study the problem of co-migrating a set of service replicas residing on one or more redundant virtual servers in clouds in order to satisfy a sequence of mobile batch-request demands in a cost effective way. With such a migration, we can not only reduce the service access latency for end users but also minimize the network costs for service providers. The co-migration can be achieved at the cost of bulk-data transfer and increases the overall monetary costs for the service providers. To gain the benefits of service migration while minimizing the overall costs, we propose a co-migration algorithm Migk for multiple servers, each hosting a service replicas. Migk is a randomized algorithm with a competitive cost of O() to migrate κ services in a static n-node network where γ is the maximal ratio of the migration costs between any pair of neighbor nodes in the network, and where λ and µ represent the maximum wired transmission cost and the wireless link cost respectively. For comparison, we also study this problem in its static off-line form by proposing a parallel dynamic programming (hereafter DP) based algorithm that integrates the branch&bound strategy with sampling techniques in order to approximate the optimal DP results. We validate the advantage of the proposed algorithms via extensive simulation studies using various requests patterns and cloud network topologies. Our simulation results show that the proposed algorithms can effectively adapt to mobile access patterns to satisfy the service request sequences in a cost-effective way.

Service migration in MEC: an Approach Based on Deep Reinforcement Learning and Fuzzy Logic

“…where t j,k denotes the time for processing results to be sent from the source node j to the destination node k. The overall service latency for marker 2 in Fig. 2 can be described in (8).…”

“…In practice, IoT-related mobile devices usually depend on the Base Station (BS) that works with the edge computing node to obtain services. However, a study [7] by the European Telecommunications Standards Institute (ETSI) has found that the coverage of BS is often limited, which leads to increased service delay when the device leaves the corresponding coverage area [8]. Additionally, in order to provide different services, multiple virtual machines running on the same edge computing node may cause I/O interference, and the corresponding service latency also increases.…”

Transparent Computing: Development and Current Status

“…It enables various OSes to share physical resources. These OSes come as Virtual machines (VMs) [15] , and the hypervisor acts as a Virtual machine monitor (VMM) to manage these VMs and allocate hardware, memory, CPU, and disk to each VM. And the container technique is a kind of kernel lightweight OS layer virtualization technique that runs on the physical host OS.…”

Online data caching in edge computing

“…Based on the identical target function, if they did not know any knowledge about the popularities of contents, Reference 22 presented an online algorithm with the competitive ratio of for the collaborative caching model in multiple cluster collaborative systems. Considering the idea of comigration, Wang et al presented a random competition algorithm which is a parallel dynamic programming algorithm based on the combination of branch and bound strategy and sampling technique 23 . Wang et al studied a fresh homogeneous cache model, that is, the transmission cost of the content between servers and caching cost of every edge node are fixed, respectively 15 .…”

“…Considering the idea of comigration, Wang et al presented a random competition algorithm which is a parallel dynamic programming algorithm based on the combination of branch and bound strategy and sampling technique. 23 Wang et al studied a fresh homogeneous cache model, that is, the transmission cost of the content between servers and caching cost of every edge node are fixed, respectively. 15 In their paper, they presented a O(mn) polynomial time algorithm upon dynamic programming to minimize total cost, where m and n are the number of nodes and the number of requests, respectively.…”