All the benefits of videoconferencing solutions, both in terms of effectiveness for companies and in terms of functionality and work organization for end users (about which we talked in this article), are not exempt from operational or infrastructural requirements and from technical challenges in order to make them available, particularly within big companies.
If on the one hand, in addition to paid licensed tools (also in a “for business” version with increased functionality and performance) there are also free software tools available, on the other hand their support in the company may not be free from costs and collateral requirements. In particular, in organizations where the number of users is high and where massive daily use is made of these tools, to ensure their quality it may be essential to put in place infrastructural measures or acquire specific technological solutions.
Let’s start from the fact that for an optimal and quality use of video conferencing it’s required that each workstation is equipped with a webcam and audio devices with adequate characteristics, both during recording (microphone) and playback. In many cases there are also solutions that are flexible in their use (such as wireless/bluetooth or environmental peripherals for group meetings), sometimes enriched with further functions to receive manual commands (answer, dial,…), cancel background noise or even to serve as replacements for telephone or VoIP solutions.
All of this is pretty obvious. What should not be underestimated, however, is the impact that the use of these applications can have on:
- bandwidth occupation and performance of the corporate network and related WAN connectivity;
- the overall efficiency of local devices (in the use of other applications);
- the sizing and computational volume required for server environments (on cloud or centralized infrastructures) also called upon to manage these additional services.
All these applications, in fact, require real-time coding and decoding of audio-video streams, a task which can be onerous in computational terms, given that algorithms can respond simultaneously to both the need for interactivity (execution in real-time) and the need to limit the use of bandwidth (obtain significant compression).
Each video conferencing solution uses different technologies which always guarantee quality levels equal or superior to audio-video connections in HD. Take for example one of the most popular platforms in the consumer or small-business world like Google Meet. Consistent with the WebRTC standard, Google Meet uses the VP8 and VP9 codecs for the compression of its video streams, and instead uses the Opus audio codec for the compression of the voice stream. Microsoft Teams, very widespread in the corporate environment, however, is based on specific solutions from Microsoft, which take advantage of the MNP24 protocol already implemented starting from the consumer versions of Skype. Cisco WebEx uses proprietary solutions as well as the market leader in the class, Zoom, which uses its own modified version of the SVC protocol (Scalable Video Coding, a name that represents the extension of the H.264/MPEG-4 AVC protocol). This is the same family of protocols widely used in other areas of video streaming, in particular in the world of desktop virtualization infrastructures (VDI).
In any case, what these solutions have in common is the need to allocate processing resources (in practical terms CPU -or GPU- and RAM) to the encoding and decoding of audio and video streams recorded locally (video of the end user connected and its voice trace) and received by the server that coordinates the video conference (that is, the joint audio and video stream of the other participants in the same meeting). Furthermore, these resources can increase in specific cases, for example if you want to make very high resolution videoconferences, if you use graphic effects (e.g. blurring or replacing the background), if you combine several video streams (e.g. sharing the webcam and also live sharing of your desktop/work environment), etc.
All this computational effort is made with a view to optimizing the use of network resources (busy bandwidth) also taking into account that the user often needs to use other applications which require a connection, that there could be several video conferences at the same time from the same network and that the connection could take place in any context. Consider, for example, the case in which, on the same single limited home line, there’s who work by collaborating remotely with these tools while their children there’s who’s following lessons in streaming. In the most typical case, in the corporate environment, networks are connected via dedicated high-capacity lines but there could be multiple, tens or hundreds of distinct videoconference sessions in progress, each of which generates internet traffic both in download and upload, in which domestic or smaller lines are often more limited.
Various studies and empirical tests confirm that, on average, the various video conferencing applications can be managed optimally in the case of a single session with a downstream bandwidth availability of 2Mbps, although this may be even lower in some cases. Similarly, a single session requires an upstream traffic, mainly of the UDP type (therefore non-synchronous sending and with the possibility of package loss), of about 1.5 Mbps, with the possibility of optimization in the case of simultaneous and concurrent connections (e.g. 10 concurrent sessions can remain within the limit of 5.0 Mbps). In fact, all the solutions try to manage the traffic generated by multiple sessions or involving multiple “peers” within the same network with specific optimizations.
What does all this teach? That the use of collaboration and video conferencing tools must also be suitably accompanied in terms of software, hardware and infrastructure, so that it’s effective and functional. If some points of attention already arise for the single user connected by a remote line, all the more these factors come into play for the corporate infrastructures (which must support the overall/cumulative load of all the simultaneous sessions carried out by their staff). In fact, we’re talking about an application area, based on communication and human interaction (transmission of voices, faces, video interaction,…) where performance is fundamental and any loss of communication makes the user experience and interaction unsustainable, therefore effectively inhibits the use of the tool.
Therefore, there are some requirements: first of all, tested and certified hardware devices for this use case and adequate and appropriately sized network connectivity, stability and bandwidth on the user base and on the use made of these tools (sporadic, frequent or massive) in combination with other commonly used and potentially “CPU or network-intensive” services. However, when it comes to the use of video collaboration in the company, the software configuration of the server infrastructure is also crucial. We’ll soon see why.
