Exam 300-820 topic 1 question 129 discussion

Local reachable Video Mesh resources are tried first, in order of lowest SRT delay. When all local resources are exhausted, the participant connects to the cloud. While the preference for node selection is your locally deployed Video Mesh nodes, we support a scenario where, if the STUN round-trip (SRT) delay to an on-premises Video Mesh cluster exceeds the tolerable round-trip delay of 250 ms (which usually happens if the on-premises cluster is configured in a different continent), then the system selects the closest cloud media node in that geography instead of a Video Mesh node. https://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cloudCollaboration/wbxt/videomesh/cmgt_b_webex-video-mesh-deployment-guide/cmgt_b_hybrid-media-deployment-guide_chapter_010.html

correct Ol_Mykhailiuk ... forgot about the 250 ms SRT limit for on-prem cluster

Cloud Cluster Selection for Overflow Based on 250 ms or Higher STUN Round-Trip Delay https://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cloudCollaboration/wbxt/videomesh/cmgt_b_webex-video-mesh-deployment-guide/cmgt_b_hybrid-media-deployment-guide_chapter_010.html

Answer should be B. Clusters in Video Mesh You deploy Video Mesh nodes in clusters. A cluster defines Video Mesh nodes with similar attributes, such as network proximity. Participants use a particular cluster or the cloud, depending on the following conditions: A client on a corporate network that can reach an on-premises cluster connects to it—the primary preference for clients on the corporate network. Clients joining a Video Mesh private meeting only connect to on-premise clusters. You can create a separate cluster specifically for these private meetings. A client that can't reach an on-premises cluster connects to the cloud—the case for a mobile device unconnected to the corporate network. The chosen cluster also depends on latency, rather than just location. For example, a cloud cluster with lower STUN round-trip (SRT) delay than a Video Mesh cluster may be a better candidate for the meeting. This logic prevents a user from landing on a geographically far cluster with a high SRT delay.