Management Plane
As mentioned previously, network devices of the past were managed individually via the CLI. Cisco Catalyst SD-WAN, however, introduces SD-WAN Manager (formerly vManage), which is a network management system (NMS) that provides a single pane of glass to manage Catalyst SD-WAN. SD-WAN Manager can be used for device onboarding, provisioning, policy creation, software management, troubleshooting, and monitoring.
While SD-WAN Manager offers a rich feature set, if the preference is to interface with it programmatically, SD-WAN Manager also supports communication via REST APIs. In fact, the SD-WAN Manager GUI is fully API driven, meaning that actions performed in it are executed using REST API calls. With full access to SD-WAN APIs, users can automate tasks, build scripts, and interface with SD-WAN Manager programmatically.
As you can see in Figure 2-7, SD-WAN Manager provides an intuitive and easy-to-consume dashboard. When you first log in to SD-WAN Manager, you are presented with an overview of the current state of the network.
)
Figure 2-7 Cisco SD-WAN Manager
vManage deployment options range from standalone nodes to three- or six-node clustered setups, offering enhanced scale and redundancy. A single SD-WAN Manager can potentially handle up to 1000 to 1500 devices, and a six-node SD-WAN Manager cluster may support more than 10,000 devices. It’s important to note that these numbers may vary based on a number of factors, such as SD-WAN Manager resources (instances/CPU/RAM/storage), the statistics load, and the version of SD-WAN software in use. (Numbers mentioned in this chapter are specific to Version 20.12.) For more accurate specifications, please consult the “Recommended Computing Resources for Cisco Catalyst SD-WAN Control Components” document for the SD-WAN software version you are using or are planning to use, available on the Cisco website.
An SD-WAN Manager cluster is designed to tolerate the failure of a single server, but for high availability, a standby cluster should be implemented to handle a complete cluster failure. Typically, it is deployed in a geographically redundant location, such as a secondary data center in another region.
SD-WAN Manager can use multiple authentication sources, including RADIUS, TACACS, and SAML 2.0, for external user connectivity. By default, SD-WAN Manager is deployed in a single-tenant mode; however, if the requirements call for support of a service provider model, multi-tenancy may be used.
All configuration for the SD-WAN fabric should be performed via SD-WAN Manager in order to maintain consistency and scalability. As discussed further in Chapter 4, you can build device configurations in SD-WAN Manager via configuration groups, feature templates, or CLI templates. You can also configure policies to control things such as network topology, routing, QoS, and security in SD-WAN Manager. SD-WAN Manager is also where you perform troubleshooting and monitoring of the network. Network administrators can simulate traffic flows to show data paths, troubleshoot WAN impairment, analyze traffic flows in the network with Network-Wide Path Insights (NWPI), and access real-time operational information (such as routing tables) for all network devices. This greatly simplifies operations as there is no longer a need to log in to each WAN Edge router individually. Instead, troubleshooting can be accomplished via a single dashboard.
Each WAN Edge router forms a single management plane connection to SD-WAN Manager. If a device has multiple transports available, only one will be used for management plane connectivity to SD-WAN Manager. If a cluster is in place, the control connection will be load balanced across cluster nodes. If a transport hosting the management plane connection experiences an outage, the WAN Edge router will briefly lose connectivity to SD-WAN Manager, and any changes made will be pushed when the device reconnects.
The last component in the management plane is SD-WAN Analytics (formerly vAnalytics). As shown in Figure 2-8, SD-WAN Analytics gives the network administrator predictive analytics to provide actionable insight into the WAN. With SD-WAN Analytics, the business can perform trending and capacity planning of circuits, and it can review how application performance is trending globally. With capacity planning, you can see how new applications may interact on your WAN before actually deploying them, allowing your business to right-size connectivity. SD-WAN Analytics ingests data from the network and uses machine learning to predict capacity trends. SD-WAN Analytics is cloud based, it requires additional licensing, and it is not enabled by default.
)
Figure 2-8 Cisco SD-WAN Analytics