My personal S3 univirse

Автор: | 19 октября, 2022

Yeah, well, I’m gonna go build my own theme park, with blackjack and hookers. In fact, forget the park! (c)

Disclaimer — about 70% of this work was made by myself. Except — Ceph clusters, Nginx VMs and Monitoring VM.

This lab is located on my work (Jet Infosystems) in VMware , consist of some Enterprise and Opensource S3 storage solution (HCP, Ceph, Minio) and it also includes some other layers to emulate production environment.

The main purpose of this environment was to demonstrate different S3 solutions to our customers, provide Proof Of Concept, convince in our competencies and show hard skills.

Application Layer — group of VMs with installed different types of S3 clients (mc, s3cmd, awscli) and s3tester for generate production workloads. At the top of the screen you can see the Service Layer — its VM with monitoring(Grafana + Prometheus) and Active Directory based on Windows Server. It was my first VM in lab at all and now it used like local DNS and help to demonstrate domain authentication for Hitachi HCP.

In the middle of the screen you can see Load balancing Layer. Based on my experience not all vendors has internal or embedded load-balancing tools for input and output S3 operation, although most of them recommend to use it. So I decide to use HAprosy and Nginx to evenly spread workload between nodes.

In the bottom of screen you can see Storage Layer. It include:

  • two Hitachi HCP clusters (with 4 VMs as a fron-end node, dedicated frontend and backend networks in each cluster)
  • two Minio clusters (with 4 VMs in each cluster)
  • two Ceph clusters ( with 3 VMs in each cluster)

All clusters have a replication based on their own functionality. They also have versioning, quotas, multipart upload, tiering to public cloud and object expiration functionalities.

Demo of S3 solutions

Добавить комментарий

Ваш адрес email не будет опубликован. Обязательные поля помечены *