June 2020, in response to multiple inquiries from a global broadcast market reacting to a global pandemic, Telos Alliance® transformed the Telos Infinity IP Intercom Platform (Infinity IP) from a hardware product into a working Cloud deployable software solution. The software team had a working beta available for testing within just a few months from a standing start. The team proved that the physical panel firmware developed for the hardware product launched in 2018 could be deployed using containerized software technology and replicate the functionality of the hardware platform almost entirely.
As with most development projects, the old adage of completing 90% of the work in 10% of the time and the remaining 10% taking another 90% has been true of Telos Infinity Virtual Intercom Platform (VIP). The many variations in use cases, internet security practices, and even upgrades to mobile phone platforms influenced the project timeline while providing a valuable learning experience for the team that will streamline many future Telos products.
Infinity VIP has been met with universal praise as an innovative solution that meets and anticipates market demands: Here is the story behind it.
Intercom is a mission critical part of most broadcast workflows, and as TV broadcast is transitioning towards a virtualized model, it seems only logical that intercom should be virtualized as well. This is the stance that Telos Alliance took back in mid-2020 when realizing that although the hardware-based Infinity IP supported remote working models (one of the driving forces behind an accelerated move towards the virtualization of video production), it simply wasn’t practical to package and ship a physical panel to every remote participant, and then in addition, hope that the recipient was tech-savvy enough to get the device online at their location.
Telos always planned to develop a software panel add-on to the Infinity system; however, with the emergence of Cloud workflows and a global pandemic accelerating the need for more agile communications solutions, a software panel solution became the number one priority in meeting the needs of a broadcast industry during an incredibly challenging time.
So how do you take a product built on a hardware platform and turn it into a software equivalent? More importantly, how do you do it quickly and from a standing start?
The legacy Telos Infinity hardware platform is a pure Audio over IP (AoIP) solution with a distributed architecture. There is no central engine or matrix to replicate, and a system can be deployed on any AES67 compliant network. Each device is ‘standalone’ and does not rely on others to function, only to communicate between one another. Given this architecture, a virtualized version of the product could theoretically be designed around the same principle, with individual software panels replicating their hardware equivalents in some way. But how could this become a reality? Enter software containerization.
A container is a lightweight means to package software code and all its dependencies so that an application can run quickly and reliably within a virtualized computing environment. Docker is an example of containerization technology and already familiar to the Telos software team. They determined the core operating code developed for the physical Infinity product could be containerized using Docker to create a virtualized network of individual devices (panels) that could communicate to one another using the same core AoIP network topology, if they were connected on a common VLAN which would act like an AES67 media network. In short, a virtual intercom platform could become a reality using Docker and operate in much the same way as physical Infinity hardware. Telos Infinity VIP was born.
Creating a network of virtual intercom panels that can communicate with one another is one thing but transforming it into a product that could enable users anywhere in the world to communicate with each other is something else entirely. But the Infinity virtual panels used the same core code as their hardware siblings which in turn already had a built-in web server, supported device discovery using Telos’s own Livewire protocol and could receive and transmit AES67 audio using common intercom features like Partylines, Groups, IFBs and Peer to Peer calls. The answer came down to using standard media protocols supported by the web server, which would in turn, be used by anybody with a compliant browser. If users navigate their browser to connect to the web server of a virtual panel, they would be looking at a panel on their screen and they could control it with the click of a mouse. If the audio devices on that computing device are supported by the browser, they would have an input and an output on their virtual panel using web audio. Plug in a USB headset and voila, virtual intercom! It sounds easy, and to an extent it was relatively simple to get that far, but that’s where the 10% / 90% rule kicks in. All the tricky stuff.
Web audio being used for intercom is great provided your computer isn’t doing much else. But this is a product that is going to be used in virtualized media production, so the computer could be busy with a multitude of other tasks and won’t care about audio packets arriving on time and in order. Web audio doesn’t do well in a busy environment, but WebRTC does. WebRTC is designed to be robust, uses the Opus codec, supports variable bitrates, and has in-built encryption. The team moved onto making WebRTC the default connection mechanism for VIP.
Understandably, intercom users want secure connectivity and reliable audio. WebRTC is perfect for this too, but when traversing the Internet, it needs Stun and Turn to make robust peer-to-peer connections, so that capability was added to VIP. Many customers don’t want to use public Stun and Turn servers because they may expose the system to security risks; therefore, the team had to address that concern. And software products need to be managed by a licensing mechanism. Additionally, there was a need for a metering system to support Software as a Service (SaaS) use cases, private DNS, a secure invitation mechanism so the right person gets to use a panel through an auto-generated password, and finally, API support for 3rd party control devices like Elgato Stream Deck.
Except, it isn’t really ‘and finally’. There was also the need to build a virtualized four-wire interface so that VIP can connect to non-panel-based audio (like Mixer Auxes, Mix Minuses, Camera Comms, Talent etc.), plus a phone App to create a better user experience for customers with Android and IOS devices, NDI support (hopefully both available by the time you read this piece) and a seemingly endless stream of feature requests from the growing number of interested customers for whom VIP seems to have really struck a chord.
Roughly 18 months after development started, VIP is used around the world with customers who appreciate the long hours that went into building it. It is a genuinely one-of-a-kind product and can be considered a disruptor. Here at Telos, we are very proud of VIP and the dedicated, brilliant team of Telos design engineers who continue to push the envelope.
