Organizations leveraging Portworx have discovered how simple, reliable and secure it is to run Kubernetes-based stateful applications in production. Underneath this simplicity, however, there is a great degree of flexibility in how Portworx can be deployed. These deployment options, or architectures, vary along two primary dimensions:
- The degree to which they address application requirements around performance, elasticity, disaster recovery, and high availability
- Their ability to operate under different infrastructure constraints such as the number of data centers and network latency.
This section outlines the following architectures:
Single data center/multiple-AZ Portworx deployment options
play_circle_outline Option 1- Dedicated Portworx storage cluster
play_circle_outline Option 2- Hyperconverged Portworx storage cluster
DR and multi-site HA architectures
play_circle_outline Option 1- Synchronous DR over a MAN using multiple Kubernetes clusters with a single Portworx stretch cluster
play_circle_outline Option 2- Asynchronous DR over a WAN using multiple Kubernetes clusters with multiple Portworx clusters
play_circle_outline Option 3- Multi-site data center stretch cluster for HA
Some of these deployment architectures are designed to be used together to solve multiple needs at once. You can pick one of the single data center Portworx deployment options and one of the disaster recovery (DR) and multi-site high availability (HA) architectures based on your requirements. For example, organizations can use the “synchronous Portworx-DR” architecture along with the “hyperconverged” architecture to achieve DR for high-performance applications.