You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The Rack Aware Cluster architecture is a hyper-converged infrastructure (HCI) solution that provides high availability and disaster recovery capabilities. It is designed to support workloads that require low latency and high throughput, making it suitable for various applications, including databases, analytics, and virtualization. <!--check if good-->
18
22
19
23
This article describes the network design and configuration of the Rack Aware Cluster architecture on Azure Local. The configuration involves a single cluster where the nodes are placed in different physical locations within a building. The intent is to support disaster recovery scenarios by placing different workloads in one or both locations. This configuration can support environments with or without Software Defined Networking (SDN), Layer 2 virtual networks, or Azure Kubernetes Service (AKS).
20
24
@@ -26,7 +30,7 @@ At the top of the setup there are two spine devices that span both rooms support
26
30
27
31
The designs illustrate the use of a single compute intent, but can support multiple intents depending on the scenario. The Storage Spaces Direct virtual LANs (711 and 712) aren't hosted at the spine layer; rather they are only hosted at the TOR layer. The Storage Spaces Direct VLANs are Layer 2 broadcast domains, and the switch doesn't support IP in the VLAN configuration. Depending on the specific Azure Local deployment scenario, the spine switches can support configurations such as SDN and AKS.
28
32
29
-
The TOR switches are Layer 2 switches that have the following configuration paths: spine uplinks for the compute and management traffic, and storage for room-to-room Storage Spaces Direct communication. The Storage Spaces Direct traffic is not distributed to the spine. Each TOR switch has uplink connections to spine devices to distribute the compute and management intent traffic. The Storage Spaces Direct traffic is isolated on specific TOR switches.
33
+
The TOR switches are Layer 2 switches that have the following configuration paths: spine uplinks for the compute and management traffic, and storage for room-to-room Storage Spaces Direct communication. The Storage Spaces Direct traffic is not distributed to the spine. Each TOR switch has uplink connections to spine devices to distribute the compute and management intent traffic. The Storage Spaces Direct traffic is isolated on specific TOR switches.
0 commit comments