Core Concepts

This section provides an overview of the Zoom Workplace VDI app's Core Concepts.

Plug-In Optimization Modes

The Zoom Workplace VDI app supports three modes of operation for real-time media processing: Direct Optimized, Channel Optimized, and Fallback Mode

In the context of the Zoom Workplace VDI app, real-time media processing refers to the relay and rendering of real-time media between the Zoom Cloud, the Zoom Workplace VDI app, and/or the Plug-In. To support a range of VDI use cases, the Zoom Workplace VDI app supports three distinct modes of operation for media processing and optimization: Direct Optimized mode, Channel Optimized mode, and Fallback mode. These are discussed in the following sections.

Direct Optimized Mode: When the Zoom Workplace VDI app and Plug-In receive independent data streams from the Zoom Cloud

Direct Optimized Mode is the default optimization mode for the Zoom Workplace VDI app and Plugin. In this mode, the Zoom Cloud maintains two separate data streams for an optimized VDI user: one for the Zoom Workplace VDI app and another for the Plug-In. This configuration enables the user’s remote client (equipped with the VDI Plug-In) to communicate directly with the Zoom Cloud for real-time media data transfers, eliminating the need to route most real-time media traffic through the virtual desktop or across the virtual channel.

When operating in Direct Optimized mode, the following occurs:

1. The Plug-In receives data streams for video and audio directly from the cloud.

2. The Zoom Workplace VDI app handles general meeting data, such as participant information, chat message, or AI Companion features, displaying it within the Workplace app placeholder, while also managing inbound screen sharing by forwarding it to the plug-in and uploading local screen sharing content from the virtual desktop when active.

3. The Plug-In and VDI desktop use the VDI vendor’s virtual connection to communicate and determine the placement and rendering of on-screen media between the two layers.

Diagram illustrating how the Zoom cloud transmits data to two separate destination when using direct optimized mode.

Channel Optimized Mode: When the Plug-In receives data hairpinned through the virtual desktop

Channel optimization is similar to the Direct Optimization experience, where the Plug-In continues to render the meeting media (as seen in the above image), but through a different network path. In this mode, the following occurs:

1. All meeting media is first delivered to the VDI server from the Zoom Cloud.

2. The VDI server transfers media to the Plug-In either through an out-of-band UDP connection or through the existing VDI virtual channel if the UDP connection cannot be established.

This method may be preferred by organizations that do not enable direct Internet access for thin clients (or other remote devices), or who prefer to route data through their network, but can potentially lead to a worse experience than the Direct Optimization if network routing conditions are sub-optimal. The image below demonstrates the UDP/Channel optimization dataflow.

Diagram illustrating how data is transmitted to the VDI Desktop and the Remote Client through a hairpinned connection.

Fallback Mode: When all meeting media is routed to and processed directly on the virtual desktop

Fallback Mode represents a fully unoptimized VDI experience. In this mode, there is no media optimization or Plug-In in use, and all communication occurs directly between the VDI server and the Zoom Cloud, with all processing occurring exclusively on the VDI server.

This method places a significant processing burden on VDI server resources, often resulting in poor performance, including slowness, choppy video, and distorted audio. As such, Fallback Mode is the least preferred option and should only be used as a last resort or when Plug-Ins are unavailable.

Warning

Fallback Mode should be avoided when possible to maintain server performance.

Summary of Connection Modes

The Zoom Workplace VDI app supports three distinct connection modes, each tailored to different operational and security needs. The default and most efficient mode is Direct Optimized Mode, where the Zoom Workplace VDI app and Plug-In establish separate connections to the Zoom Cloud, independently handling their respective portions of a Zoom meeting to deliver a seamless, optimized experience.

In addition to Direct Optimized Mode, the Zoom Workplace VDI app can operate in alternative configurations, including Channel Optimized Mode and Fallback Mode. These modes can help address specific workflow or network constraints, such as restricted internet access for remote devices, data routing for privacy concerns, or the absence of Plug-Ins.

The following table summarizes the key differences between these modes.

	Media offloading	Direct cloud access from Plug-In
Direct Optimized	✔	✔
Channel Optimized	✔
Fallback Mode

WebRTC Media Offloading

Overview

Zoom provides a browser-based WebRTC client via the Zoom Web App that can offload audio processing to the user’s local device when running inside a virtual desktop environment. This works without requiring any Zoom-specific plug-ins because the VDI platform supplies its own local WebRTC engine and a redirection framework that bridges the Zoom Web App to that engine.

Warning

WebRTC Media Offloading is currently limited to audio and does not support video optimization.

This feature supports the following products and channels from the Zoom Web App:

• Zoom Phone

• Zoom Contact Center

• Zoom Contact Center CTI Connector

This feature is currently supported by the following virtual desktop platforms:

• Citrix

• Omnissa Horizon

Refer to Zoom’s support center for more information on configuring Zoom VDI to support WebRTC redirection for the Zoom Web App.

The Zoom Web App Offloads VDI WebRTC Audio to the Local Device

When the Zoom Web App inside the virtual desktop attempts to initialize WebRTC audio, its requests are intercepted before the virtual desktop tries to capture or process sound. Instead of activating the browser’s built-in WebRTC media stack in the hosted session, the VDI platform translates the audio-related signaling into lightweight control messages. These messages are sent through the VDI provider’s virtual channel to the user’s local machine, redirecting real-time audio traffic from the Zoom cloud directly to the user’s local machine.

On the local machine, the native WebRTC engine included with the VDI client (e.g., Citrix, Omnissa Horizon) receives these messages and becomes responsible for all audio capture, encoding, decoding, and playback. The engine uses the local system’s microphone, speakers, and processing resources, helping ensure that audio does not flow through the virtual desktop server.

The following diagram illustrates how data is routed when using WebRTC Media Offloading with the Zoom Web App and a supported virtual agent.

Diagram illustrating how media is split between the Virtual Desktop and Remote Client device.

Interaction Between the Zoom Web App and the Local Machine

From the Zoom Web App’s perspective, the experience still resembles a standard WebRTC session. Signaling between the Zoom Web App and Zoom’s backend is relayed through the virtual desktop, and the local WebRTC engine mirrors the negotiated session parameters. The virtual desktop application continues to present the Zoom interface—controls, meeting status, and indicators—while the actual real-time audio is generated and consumed by the local machine.

Because only signaling messages traverse the virtual channel, bandwidth overhead is low and consistent, even in multi-user environments.

Why No Plug-In Is Required

The key enabler is that the VDI client (e.g., Citrix, Omnissa Horizon) already includes a full WebRTC media engine capable of handling real-time audio. Since the redirection layer makes this engine appear to the Zoom Web App as its underlying WebRTC implementation, Zoom does not need to provide and maintain a separate plug-in. The redirection logic maps WebRTC API calls, device access, and session negotiation from the browser inside the virtual desktop to the native engine on the local machine.

Result

This approach allows Zoom’s browser-based WebRTC experience to operate efficiently in VDI environments with full audio optimization. The interface runs inside the virtual desktop, but real-time audio is captured and processed locally, giving users a responsive and scalable conferencing experience without requiring additional Zoom software on the local machine or processing demands from the virtual desktop’s server.