GigaOm Benchmark: New Teams Optimization for Virtual Desktop Infrastructure (VDI)v1.0

1. Executive Summary

This GigaOm Benchmark was commissioned by Microsoft.

Microsoft Teams is a collaboration platform that combines workplace chat, video meetings, file storage, and application integration. Microsoft Teams enables users to communicate and collaborate in real-time, share files, and access applications from a single platform.

Last year, GigaOm published a benchmark that compared the performance and behavior of the previous version of Teams (“Classic Teams,” version 1.x) against the recently released version of Teams (“New Teams,” version 2.x). That benchmark showed massive improvements in administration time, responsiveness, resource utilization, and overall performance for New Teams. Much of the improvement in the user experience was due to Microsoft’s decision to shift to a more responsive Microsoft-native UI built on the Microsoft Edge WebView2 framework.

In this benchmark, we explore the experience running the new Microsoft Teams collaboration platform in a virtual desktop interface (VDI) environment powered by either Citrix Desktop or Azure Virtual Desktop (AVD).

This GigaOm benchmark report coincides with Microsoft releasing a new optimization stack for Teams VDI, which uses Microsoft-native Slimcore based optimizations. Until now, both Classic Teams and New Teams have used a WebRTC redirection-based optimization, which requires a separate service running on the VDI host server. The new Slimcore optimization is built directly into Teams and is supported by Azure Virtual Desktop, Windows 365, and Citrix.

Importantly, the new Teams optimization on VDI does away with WebRTC helper applications on VDI host servers, which add administration time and complexity to update and manage in the VDI environment. The new optimizations yield valuable performance and efficiency gains for both the client and the server—enabling more user-density per host and improving user experience through a more responsive user interface, faster meeting join times, and higher quality multimedia content such as screen sharing and video.

Across all tests, including both Azure Virtual Desktop and Citrix DaaS, New Teams produced a 25.2% average reduction in server-side CPU usage compared to Classic Teams, and a 51% reduction in average memory usage per user. Figure 1 illustrates the relative improvement produced by New Teams over Classic Teams in each test. Note that these metrics are averaged across multiple calls, configurations, and server utilization throughout each call. Generally, the more multimedia content, the more effective the new optimization is in comparison to the outgoing stack.

Figure 1: Relative Server Resource Utilization Compared to Classic Teams

Figure 1 shows the significant reduction in server-side resource utilization per user, enabling more user density per session host and directly reducing the costs of a large-scale VDI deployment.

Beyond the feature and performance improvements, upgrading to New Teams for VDI provides streamlined administration. The new optimization stack is built directly into the Teams application, decoupling the solution from Citrix Visual Delivery Agent (VDA) and Remote Desktop (RD) Host versions/services.

In addition, the Microsoft RD client and Windows App bundle the required new components out of the box, and the Citrix Workspace app has multiple delivery mechanisms for the components (Global App Config Service, command line installer arguments, UI option, or Intune/SCCM).

2. Introduction

Virtual desktops are not new. VDI applications and virtual desktops have evolved over time, with commercial applications dating back as early as the 1990s.

What is new is the ability to run highly responsive and media-rich workloads in VDI environments—for example, live video conferencing applications. Fast and reliable internet connections, advanced media encoding, and robust streaming technologies have enabled efficient, low-latency video streaming. Companies like Microsoft have invested resources into developing and optimizing the technologies that enable these solutions.

New Teams on VDI demonstrates a like-local user experience, with high-quality video, a responsive user interface, and a full suite of features comparable to those provided to Teams running on a local device. The new Teams optimization on VDI architecture adds to these improvements with more efficient resource utilization on both client and server operating systems, support for more video streams, the latest codecs, and more.

3. Field Test Overview

To test the new Teams for VDI, we deployed two environments to assess the new optimizations running on both Microsoft AVD and Citrix Desktop.

We conducted test calls using Classic Teams with VDI 1.0 and New Teams with the new optimizations. It’s worth noting that Classic Teams with VDI 1.0 supports just four video streams. By contrast, New Teams with the optimization has no such limitation, supporting the same number of video feeds as Teams would running locally.

To capture resource utilization, we measured the CPU and RAM usage of the Teams application and its subprocesses. For Classic Teams with VDI 1.0, the WebRTC redirector service is also included in the measurements. On the local device, we measured the VDI client (Citrix Workspace or Microsoft Remote Desktop) as well as the associated subprocesses. A detailed breakdown of each service is outlined in the appendix.

For Azure Virtual Desktop, a virtual host was deployed using an Azure VM in West US 3. The SKU selected was a D16s_v5 with 16 vCPU and 64 GB RAM. Our Citrix lab was hosted on a server with similar specs, however with a different processor SKU. The configuration of solution stack components is shown in Table 1.

Table 1. Solution Stack Version Information

Note: The Intel Xeon 8171M used in the Citrix testing is from an older processor generation than the 8370C used to test AVD. As a result, the CPU usage metrics are not directly comparable between our AVD and Citrix environments. This has no impact on the overall thesis behind this benchmark, as we are not comparing Teams on Citrix to Teams on Azure Virtual Desktop. The focus is on comparing the behavior of New Teams and Classic Teams in two distinct VDI environments.

4. Field Test Report

In our testing, we investigated the performance and behavior for VDI implementations of New Teams and Classic Teams. If your organization has already switched to New Teams on VDI, no further action is needed to update your VDI session hosts. When the new optimization stack is released, New Teams will automatically select the optimal configuration supported by the client. It is recommended to update to the latest version of Microsoft Remote Desktop or Citrix Workspace for the most up-to-date and secure connection into your VDI environment.

All of the performance improvements measured in our previous benchmark are still reflected in the server-side environment for Teams on VDI. For instance, findings like faster installation times and the ability to deliver via MSIX package make it easier than ever to upgrade to and maintain a large-scale deployment of Microsoft Teams on VDI. We found that each user’s local data footprint for New Teams was roughly half that of Classic Teams. Multiplied across hundreds or thousands of user profiles, the resulting storage savings is significant.

Note: To provide ease of comparison across tests, the charts in this section display New Teams performance relative to the measured performance of Classic Teams. As such, Classic Team is always scored at 100%, while the New Teams value shows the extent to which New Teams reduced the time or resources required to complete each of the tests.

Server Memory Utilization

A key finding for VDI is that New Teams offers impressive reductions in memory utilization, a benefit that is multiplied when run at scale. On average, New Teams slashed the per-user, 90th-percentile memory footprint by more than 500 MB. To put this into perspective, an environment hosting 50 users on Classic Teams would require 25 Gigabytes more RAM than a server with the same 50 users on New Teams.

As shown in Figure 2, New Teams uses significantly less memory across the board in our tests based on 15-minute calls, reflecting a variety of meeting scenarios. Here we see consumption reduced by 30% to 40% in tests involving video or screen sharing. Each test measures the total server-side memory consumption of the Teams application for each participating user in the test.

Figure 2. Relative Memory Utilization (GB)

Further, the 4-person video result in Figure 2 illustrates how New Teams provides increasing memory efficiency in media-rich tests when compared to Classic Teams.

Server CPU Utilization

To simplify CPU sampling, we measured the total CPU time of all Teams executables and subprocesses for the duration of each 15-minute call. This means that for a 4-user call with screen sharing, the total CPU time includes the Teams executable for the presenter as well as three observers. This mechanism does not capture the maximum instantaneous load from inconsistent workloads, nor a single-user’s maximum consumption, but it is an effective measure of the total load the Teams application puts on the server’s CPU at scale. When averaged out over dozens of users per server, the intermittent highs and lows will effectively average out to reflect the measurements we compare here in the report. Additionally, in real world environments, Teams with the new VDI optimization will consume only a portion of the overall load each user puts on a VDI host when you account for web browsers, productivity applications, and the VDI session itself.

Figure 3 shows the relative CPU consumption by New Teams compared to Classic Teams in each of our test scenarios. Across all three tests, CPU utilization falls by 20% compared to Classic Teams.

Figure 3. Relative Server CPU Usage

Client Resource Utilization

Our testing revealed that New Teams with Slimcore-based optimization produced measurable reductions in resource utilization on the client as well.

The tests below were run with Classic Teams with WebRTC-based optimization vs. New Teams with Slimcore-based optimization. However, it is worth noting that the resource consumption of the WebRTC-based optimization is independent of the Teams version (Classic or New).

Figure 4 shows the average memory usage of the VDI client. For Microsoft Remote Desktop, all resource consumption is contained within a single executable. For Citrix, we tracked a list of processes and subprocesses, including the associated client-side VDI optimization for either Classic Teams (WebRTC) or New Teams (Slimcore).

Figure 4. Relative VDI Client Memory Usage

Note: These measurements are averages across multiple test calls, and the memory allocation of the remote desktop client fluctuates based on numerous factors, including session length, webcam redirection, screen resolution, and more.

Figure 5 shows the relative client-side CPU comparison between Slimcore and WebRTC for both AVD and Citrix. Again, the results are normalized to show the performance of Slimcore relative to WebRTC, which is why WebRTC scores 100% on these tests. In this case, the reduced CPU utilization is largely due to the addition of hardware (GPU) acceleration with the new optimization stack.

Figure 5. Relative Client-Side CPU Usage

Battery Life

In our testing, we found that more efficient client-side resource utilization produced a minor reduction in battery utilization. To measure this, we extended our call length to 30 minutes and disconnected the client system from external power. We disabled Windows updates; however, we did not extensively control for other background tasks. On average, our VDI client running New Teams with the new optimizations consumed about 14% of the available charge, while Teams Classic consumed approximately 15.3% over each 30-minute call. This can be extrapolated to 214.2 minutes of battery life on New Teams, compared to 196.2 minutes of battery life on Classic Teams, a 9.17% increase in battery life. Note that this battery life increase may be less pronounced as additional workloads are run on the client machine.

Constrained Network Bandwidth Test

We assessed Classic Teams and New Teams in a constrained networking environment to determine performance and behavior in less-than-optimal conditions—a common use case for remote video and audio calling. For this portion of the evaluation, our analysts assessed the observed quality of the video during a series of test calls. We used NetEM to restrict bandwidth and induce dropped packets, which are common causes of degraded call quality. We observed the following:

• Given unlimited bandwidth, New Teams was observed to use more bandwidth than Classic Teams due to its support for higher bit rates and higher-quality video.
• At lower bandwidths, New Teams was observed to maintain superior audio quality and visual fidelity compared to Classic Teams.
• At the lowest bandwidths and highest packet loss configurations (below 100 kbps and above 20% packet loss—a configuration significantly outside supported network conditions—both New Teams and Classic Teams were unable to support video. However, New Teams was better able to recover from audio dropouts to continue playing the audio stream compared to Classic Teams.

Based on our observations (shown in Table 2), we are confident that for users with limited-bandwidth or unreliable connections, such as 4G mobile workforces, New Teams on VDI will prove noticeably more reliable and effective to use compared to Classic Teams.

Table 2. Constrained Network User Experience Comparison

Call Join Time

The time it takes to join an audio or video call may not seem critical, but load times have high-stakes psychological effects on user interface perception. Busy workers in back-to-back meetings or meeting hosts running late will appreciate the quicker join times with New Teams.

New Teams for VDI benefits from the same improvements to call join time as New Teams on your local device. We tested the experience by setting up two scenarios: A two-person audio call where we measured the time to join for the second participant joining, and a four-person screen sharing call where we measured the fourth participant joining. Time was measured from the point the last participant clicked to enter the call until all videos on the call were fully rendered.

As shown in Figure 6, users can expect to experience up to a 30% reduction in call join times with New Teams, based on the four-person test. The time savings for the two-person audio call was less, but still significant at 16%. With larger meetings and more active video feeds, the join times can become even more pronounced, making New Teams even more compelling to adopt.

Figure 6. Call Join Time

5. Analyst’s Take

Testing New Teams for VDI revealed significant improvements in both performance and resource efficiency, enabling higher user density on servers that can yield lower hardware costs and simplified IT management. The responsive user interface enhances productivity and improves overall user experience, allowing workers to collaborate more efficiently and maximize productive time. Overall, New Teams for VDI delivers cost savings, operational efficiencies, and a better user experience, driving business growth and improving connection among remote workers.

The benefits of the New Teams client shine through in a VDI environment. Our previous benchmark showed that New Teams produced a 2x improvement in response times, a 50% overall decrease in disk and memory usage, and a significant reduction in meeting join time and application launch time.

These benefits are compounded by the superior performance and resource efficiency of the new optimization. New Teams also provides increased battery life and higher performance for lower-end and low-power devices when running a VDI workspace instead of the whole application suite a modern worker requires, which may include collaboration tooling, productivity software, email clients, web browsers, and development environments. We noticed a measurable increase in battery life when switching from Classic Teams on VDI 1.0 to New Teams and the new optimization on VDI.

VDI enables nearly unlimited scalability in compute, memory, and storage for users, and as we found in our recent benchmark of Microsoft Dev Box, Microsoft is creating purpose-built VDI solutions for specific business use cases, such as dedicated cloud desktops for software developers and engineers. You can read our benchmark of Microsoft Dev Box here and see more about the product on Microsoft’s website here.

6. Appendix

Measured Executables

Table 3. Memory Consumption by Executable

Test Data

Here we present CPU and memory utilization test results across our three call profiles: 4-person call with screen sharing, 4-person call with video, and 2-person call with video. These tests show the side-by-side behavior of New Teams with the new optimization on VDI and Classic Teams with VDI 1.0. The tests were run in the two primary virtual desktop environments—Microsoft AVD and Citrix Desktop.

4-Person Call with Screen Sharing

Figure 7. 4-Person Call with Screen Sharing: Server-Side CPU Utilization in Seconds

Figure 8. 4-Person Call with Screen Sharing: Server-Side Memory Utilization in GB

4-Person Call with Video

Figure 9. 4-Person Video Call: Server-Side CPU Utilization in Seconds

Figure 10. 4-Person Video Call: Server-Side Memory Utilization in GB

2-Person Call with Video

Figure 11. 2-Person Call with Video: Server-Side CPU Utilization in Seconds

Figure 12. 2-Person Call with Video: Server-Side Memory Utilization in GB

7. About Eric Phenix

Eric Phenix is Engineering Manager at GigaOm and responsible for our cloud platforms and guiding the engineering behind our research. He has worked as a senior consultant for Amazon Web Services, where he consulted for and designed both systems and teams for over 20 Fortune 1000 enterprises; and as cloud architect for BP, where he helped BPX Energy migrate their process control network from on-premises to AWS, creating the first 100% public cloud control network, operating over $10 billion in energy assets in the Permian Basin.

8. About GigaOm

GigaOm provides technical, operational, and business advice for IT’s strategic digital enterprise and business initiatives. Enterprise business leaders, CIOs, and technology organizations partner with GigaOm for practical, actionable, strategic, and visionary advice for modernizing and transforming their business. GigaOm’s advice empowers enterprises to successfully compete in an increasingly complicated business atmosphere that requires a solid understanding of constantly changing customer demands.

GigaOm works directly with enterprises both inside and outside of the IT organization to apply proven research and methodologies designed to avoid pitfalls and roadblocks while balancing risk and innovation. Research methodologies include but are not limited to adoption and benchmarking surveys, use cases, interviews, ROI/TCO, market landscapes, strategic trends, and technical benchmarks. Our analysts possess 20+ years of experience advising a spectrum of clients from early adopters to mainstream enterprises.

GigaOm’s perspective is that of the unbiased enterprise practitioner. Through this perspective, GigaOm connects with engaged and loyal subscribers on a deep and meaningful level.

9. Copyright

© Knowingly, Inc. 2025 "GigaOm Benchmark: New Teams Optimization for Virtual Desktop Infrastructure (VDI)" is a trademark of Knowingly, Inc. For permission to reproduce this report, please contact sales@gigaom.com.