1 of 6

Decentralized control in large-scale

fog computing infrastructures

Chih-Kai Huang

Univ Rennes, IRISA, CNRS, Inria

Supervised by

Prof. Guillaume Pierre (Univ Rennes, Inria, CNRS, IRISA)

EuroSys Doctoral Workshop 2022

2 of 6

2

Fog computing platform has a very different geographical distribution.

Kubernetes (K8s)

is a well-known open-source platform for container orchestration.

Introduction

Fog computing platform is closer to the end-user and can serve them as quickly as possible [1].

In addition, Kubernetes is suited to be the management system for fog computing platform [2].

EuroSys Doctoral Workshop 2022

To reduce the end-to-end latency from IoT devices to the centralized data center, the concept of fog computing [4] is coming up. Fog computing platform is closer to the end-user and can serve them as quickly as possible especially for some IoT applications, which require a short response time.

Therefore, fog computing platform has a very different geographical distribution that is deployed fog nodes in a different places.

Kubernetes (K8s) is a well-known open-source platform for container orchestration. In addition, Kubernetes is suited to be the management system for fog computing platform. However,

a single Kubernetes cluster may reach the max number of resources when the cluster is large enough.

Multiple Kubernetes clusters deployment is one of the answers to solve the problems.

3 of 6

3

Kubernetes Cluster Federation (KubeFed) [3]

empower users to manage the multiple Kubernetes clusters from the one main cluster.

Related Work

It extend the KubeFed and provide users to manage the multiple Kubernetes clusters from the one main cluster.

multi-cluster Kubernetes (mck8s) [4]

EuroSys Doctoral Workshop 2022

4 of 6

4

Limitations

Both of works are centralized architectures, which will face the scalability problem.

Moreover, this architectures can not handle a single point of failure.

选题的意义

Problems

EuroSys Doctoral Workshop 2022

5 of 6

References

[1] Flavio Bonomi, Rodolfo Milito, Preethi Natarajan, and Jiang Zhu. 2014. Fog Computing: A Platform for Internet of Things and Analytics. Springer International Publishing, Cham, 169–186. https://doi.org/10.1007/978-3-319-05029-4_7

[2] Ali Jawad Fahs, Guillaume Pierre, and Erik Elmroth. 2020. Voilà: Tail-

Latency-Aware Fog Application Replicas Autoscaler. https://doi.org/

10.1109/MASCOTS50786.2020.9285953

[3] Multicluster Special Interest Group. 2021. Kubernetes Cluster Federation. https://github.com/kubernetes-sigs/kubefed/blob/master/ README.md.

[4] Mulugeta Ayalew Tamiru, Guillaume Pierre, Johan Tordsson, and

Erik Elmroth. 2021. mck8s: An orchestration platform for geodistributed multi-cluster environments. In 2021 International Conference on Computer Communications and Networks (ICCCN). 1–10. https://doi.org/10.1109/ICCCN52240.2021.9522318

References

6 of 6

hank you for listening

T

chih-kai.huang@irisa.fr