Aria Operations for Networks communicates with management components to collect metrics, topology data, and connectivity that are presented through various dashboards and views. Aria Operations for Networks collects metrics from various VMware and non-VMware products, including vSphere, NSX, Kubernetes, and physical network components.
Aria Operations for Networks - Logical Design
Aria Operations for Networks is deployed as a cluster called the Aria Operations for Networks Platform Cluster. This cluster processes the collected data and presents it using a dashboard. Aria Operations for Networks also uses collector nodes to collect data from the data sources, such as vCenter Server, NSX Manager, and Avi Load Balancer, and send the data to the Platform Cluster for processing.
Aria Operations for Networks is only included as part of Telco Cloud Platform Advanced, it is not a part of the Telco Cloud Platform Essentials bundle.
The platform cluster capacity is based on the deployment size of each node and the number of nodes composing the cluster. For more information about how to decide the platform brick (node) size and number of platform bricks (nodes), see the Aria Operations for Networks documentation.
The collector capacity is based on the deployment size. The amount of data sources that you can add to a specific collector is based on the documented capacity of the collector size (accounting for the number of VMs and the number of flows).
When using Aria Operations for Networks, it provides an option to enable IPFIX export on vCenter switches. While this provides additional information for communication between the elements of the Telco Cloud, IPFIX exporting can add latency issues. As such, IPFIX is not recommended for deployment on Workload Switches as this can impact the performance of user-plane functions.
![Logical Design Components for Aria Operations for Networks](images/GUID-F21D5E90-77D4-4E06-B319-0EBC7D14E2EE-low.png)
Aria Operations for Networks - Scaling
Aria operations for Networks scales in two dimensions: the platform and the collectors. The platform can only be clustered if using appliances of at least a large size. Medium deployments cannot be scaled into a cluster without first scaling up to large.
The scaling of a platform to a cluster model does not provide additional availability to the platform. Scaling must be used when maximums for a single node exceed.
The overall dimensioning for collector scaling is based on various factors including:
Number of vCenter Servers
Number of VMs
Number of flows to track
To determine when to size the platform and collector nodes, see the Aria Operations for Network Scaling deployment guide. When collecting over 10,000 VMs and 4 million active flows, the platform must be upgraded from a single node to a clustered deployment. Similar scale-up and scale-out requirements exist for the Collectors.
Aria Operation for Network Design Recommendations
Attribute |
Specification |
---|---|
Appliance Size - Platform |
Large |
Number of vCPUs (based on CPU speed) |
2.1 Ghz - 15 vCPUs, 2.3 Ghz - 14 vCPUs, 2.6 Ghz - 12 vCPUs |
Memory |
48 GB |
Disk Space |
1 TB |
Appliance Size - Collector |
Large |
Number of vCPUs (based on CPU speed) |
2.1 Ghz - 10 vCPUs, 2.3 Ghz - 9 vCPUs, 2.6 Ghz - 8vCPU |
Memory |
16 GB |
Disk Space |
200 GB |
Design Recommendation |
Design Justification |
Design Implication |
---|---|---|
Deploy a single, large Aria Operations for Network platform instance. |
A cluster is required based on scale. A single large platform node allows expansion into a cluster while providing initial monitoring capabilities The platform can be scaled into a cluster based on future growth or features. |
Each platform node must have 100% reservation for CPU and RAM. |
Deploy at least one Collector Deployment for every Workload Domain. |
Enables data collection locally on every Workload Domain |
Each Collector node must have 100% reservation for CPU and RAM. Dimensioning of Collector must be checked during Workload Domain creation/expansions. Collector nodes are not highly available, but they must be protected by vSphere HA. |
Platform and Collector nodes must be scaled as the network scales to accommodate the additional load. |
Aria Operations for Networks is based on components that can be scaled out or scaled up as required. |
Additional resources must be available. |
Aria Operation for Networks can be connected to the vCenter, NSX, Avi Load Balancer, and Log Insight data sources. |
Provides network visibility of the virtual networking. |
Requires valid service accounts for Aria for Operations components. |