Functionality

This section details additional information regarding Zoom Mesh functionality.

Parent-Child Functionality

The following sections provide additional information on parent and child relationships within a mesh network.

Zoom Mesh creates a network exclusively for users within the same account by default

By default, Zoom Mesh creates a network exclusively for users belonging to the same account. However, an account can optionally allow for authenticated, external users within the same local area network to participate in the Mesh, providing the external user’s account also allows them to connect to another account’s mesh network.

Parent and child roles are determined by the Zoom Mesh Cloud Orchestration Service

The Zoom Mesh Cloud Orchestration Service (COS) is a cloud service responsible for defining parent and child client roles within a network. The Cloud Orchestration Service considers numerous sets of data to determine viable parent clients including:

• CPU Utilization

• CPU type

• RAM utilization

• Historical mesh performance

• Device availability (On/Off)

• Device Opt-In/Out

• Operating System

Using these data sets, the COS scores a user’s results through a rubric: the higher a user’s score, the more likely the client is to be selected as parent; the lower the score, the less likely.

Parent and child client designation is dynamic and can change throughout an event

During the course of a meeting or webinar, a user’s client designation may change based on performance or demand.

For example, in an event of 100 attendees, only 10 parent clients may be required to meet the redistribution demand, with the remaining 90 users receiving redistributed streams as child clients. However, if the event increases to 200 attendees, clients that were formerly child clients may be promoted to a parent client to address the increased demand.

Alternatively, if a parent client begins to experience degraded performance, the client may be downgraded to child client status, and a new parent client will be assigned.

The parent-child client ratio is dynamically determined by each machine’s hardware and network performance

During an event, the parent-child ratio between clients is dynamically updated based on the device’s ongoing network and hardware performance.

For example, if a device is showing low hardware (CPU/RAM) utilization with a high bandwidth uplink capacity, the device is ideal for a parent-client role and may establish a parent-child relationship with several other devices. In this scenario, the parent-client may attempt to redistribute media to other devices, so long as the device is not oversubscribed or begins to degrade in performance. However, if the device’s network or hardware performance begins to decline, it may surrender its parent-client designation and move back to a child-client role.

Each child is assigned two parents

Child clients connected to a mesh network are assigned two parent clients for resiliency. If one parent client fails, the user will immediately failover to the second parent client. A new secondary parent client is then established to restore resiliency.

Users are not informed when they are connected to a mesh network

Event hosts and attendees do not receive in-client notifications when connected to a mesh network. Mesh network connections can only be seen by an admin or authorized user through the Zoom Mesh Dashboard.

VDI clients can only establish a mesh connection with other VDI clients

When using Zoom Mesh with virtual desktop infrastructure (VDI), client devices can only establish Zoom Mesh connections with other VDI clients, so long as both users are using a supported VDI Plug-In. This is due to technical limitations of the virtual infrastructure and relaying media between devices.

VDI clients do not distribute the content sharing channel in a mesh network

Due to the design of Zoom’s VDI Client, screen sharing content must be routed through the virtual machine. Consequently, VDI clients cannot distribute content sharing across a mesh network, and must continue to be distributed within the virtual infrastructure.

Zoom Mesh for Meetings Functionality

The following sections describe details of Zoom Mesh for Meetings functionality.

Mesh for Meetings Redistributes the Active Speaker’s Video

With Zoom Mesh for Meetings, bandwidth optimization is delivered by redistributing the downstream active speaker’s video between participants when four or more users are in a meeting. This feature does not apply to media uploading, and all users will continue to individually upload their audio, video, and screen sharing media to the Zoom Cloud.

Example

If there are 10 users in a meeting with Zoom Mesh enabled, two parent clients may be chosen with seven or eight child clients. When the current active speaker is talking, that user’s video stream will transmit from the Zoom Cloud to the parent clients, and will then be redistributed to children clients within the network.

In this scenario, the child clients do not receive the active speaker’s video feed from the Zoom Cloud, saving potentially up to 3+ megabits of data per user when using 1080p video.

The following image provides an example of how downstream active speaker video content is redistributed within a mesh network.

Mesh for Meetings active speaker changes do not alter parent-child relationships or video distribution bandwidth savings

While the focus of Zoom Mesh for Meetings is to redistribute the active speaker’s video feed, changes in the active speaker do not affect the relationships between parent and children clients or the bandwidth savings afforded with the video redistribution. When the active speaker switches, designated parent clients simply change the feed of the video they are redistributing.

Example

If Maurice is speaking in a meeting and then Holly begins speaking, the parent client’s redistribution feed will change from Maurice’s video feed to Holly’s.

This change does not alter which user clients are currently designated as parents or children, nor the general savings in bandwidth afforded by the video redistribution, only the redistributed video feed changes.

Bandwidth savings may vary depending on the user’s selected video layouts

Although Zoom Mesh saves on bandwidth consumption by redistributing the active speaker’s video feed, bandwidth savings may be affected depending on users chosen layouts. For instance, Gallery View supports up to 49 on-screen video feeds at once, while Speaker View shows one video feed at a time. If a user is watching a meeting in Gallery View, they will benefit by receiving the singular active speaker’s video feed through the mesh network, but they will continue to consume standard amounts of bandwidth by downloading the remaining 48 user’s video streams from the Zoom Cloud.

For the best bandwidth saving practices, users are encouraged to view meetings in as close to a webinar-style layout as possible, such as Multi-Speaker or Speaker View.

Zoom Mesh for Webinars Functionality

The following sections describe details of Zoom Mesh for Webinars functionality.

Mesh for Webinars redistributes all webinar hosts and panelists video streams within the network

Unlike Zoom Mesh for Meetings, Zoom Mesh for Webinars redistributes all webinar hosts and panelists video streams within the local network using parent-child relationships—not just the active speakers. With this design, all users engaged in the mesh network will continue to benefit from reduced external bandwidth consumption while experiencing the same video quality as others.

Mesh for Webinars benefits do not change with video layouts

Regardless of which video layouts users choose when viewing a webinar, users will continue to receive the benefits of video redistribution within the network, including Gallery View, Speaker View, or Multi-Speaker View.