mck8s: An orchestration platform for geo-distributed multi-cluster environments

Abstract

Following the adoption of cloud computing, the proliferation of cloud data centers in multiple regions, and the emergence of computing paradigms such as fog computing, there is a need for integrated and efficient management of geodistributed clusters. Geo-distributed deployments suffer from resource fragmentation, as the resources in certain locations are over-allocated while others are under-utilized. Orchestration platforms such as Kubernetes and Kubernetes Federation offer the conceptual models and building blocks that can be used to build integrated solutions that address the resource fragmentation challenge. In this work, we propose mck8s -an orchestration platform for multi-cluster applications on multiple geo-distributed Kubernetes clusters. It offers controllers that automatically place, scale, and burst multi-cluster applications across multiple geo-distributed Kubernetes clusters. mck8s allocates the requested resources to all incoming applications while making efficient use of resources. We designed mck8s to be easy to use by development and operation teams by adopting Kubernetes' design principles and manifest files. We evaluated mck8s in a geo-distributed experimental testbed in Grid'5000. Our results show that mck8s balances the resource allocation across multiple clusters and reduces the fraction of pending pods to 6% as opposed to 65% in the case of Kubernetes Federation for the same workload.

Matching-based Scheduling of Asynchronous Data Processing Workflows on the Computing Continuum

“…4) Data traffic reduction: Arkian et al [16] presented a geo-distributed stream processing model to sustain a sufficient parallel throughput among Edge devices, optimizing the network latency and resource utilization. Tamiru et al [17] designed and integrated a scheduler in the Kubernetes orchestration platform that models the workload requirements, resource capacity, and the ingress traffic to multiple computing clusters. Mehran et al [9] proposed a two-sided matching and resource allocation model for streaming applications that improves the data processing time of microservices and residual bandwidth of Cloud and Fog devices.…”

“…We integrated our customized scheduler in the Kubernetes 1.21 orchestration tool using the Python client library 17.17 [27], published in the GitHub code repository [28]. We employed the Prometheus Operator v0.45.0 to monitor the Kubernetes deployments, microservices, containers, and devices [17]. We used the KubeMQ 2.2.10 message queue platform to implement the asynchronous data exchange between microservices [29].…”

Good Shepherds Care For Their Cattle: Seamless Pod Migration in Geo-Distributed Kubernetes

“…Transparent migration for the application Transparent migration for the clients Uses unmodified application images is frequently used to prototype fog computing platforms [22], [23], [24] The RPIs use the HypriotOS Linux distribution, Kubernetes 1.16 and Docker 19.03.5.…”

Cheops, a Service to Blow Away Cloud Applications to the Edge

“…Mck8s [12] is another similar solution with geo-distribution application deployments. The primary objective being to deploy applications onto other locations including Edge sites.…”