Part of the benefit of open standards is that they enable multi-vendor interoperability. However, in a world as big and diverse as ours, it’s impossible to find one AV-over-IP technology that can work for every use case and affordably fit in every product that we might want to create. Organizing requirements into profiles is an excellent way to meet the needs of a broad market while keeping things simple and efficient. Developing clear and effective compatibility profiles within IPMX will have a significant impact on the future of the standard.
Today, AV-over-IP technology providers are able to deliver interoperability by limiting devices to the implementation of their protocol. As a side effect, which also enables interoperability, their solutions contain only the codecs and network interfaces that they choose to support. For SDVoE, that is 10GbE mixed with a lightweight proprietary codec and no compression. For almost everyone else, it has been 1GbE for transport and various codecs. In this way, SDVoE devices work with other SDVoE devices. With NDI and Dante AV, this can also be true, but both have variations that work with different codecs—more on this in a bit.
While these constraints deliver interoperability for proprietary AV-over-IP technology solutions, the situation is different for an open standard like IPMX.
The IPMX Open Standard
First, open standards inevitably mean multiple implementations of the same technology. Since proprietary solutions limit the market to a single implementation, they do not have this complication. However, this creates risk in the supply chain, which COVID-19 has demonstrated. Having multiple implementations also increases the robustness and security of systems, making them less susceptible to common bugs, and by making the bugs that do exist less attractive targets for attackers. That is why, on balance, the extra work required to support multiple implementations is easily worth it.
The second way that IPMX differs from proprietary interoperability solutions flows from the first. Because many more companies will implement the standard, it will be exposed to a wider variety of environments and use cases. This means that limiting a solution to a specific transport bandwidth or codec is impractical. There are applications that require nothing less than uncompressed video and others that require extremely efficient use of bandwidth. For some applications, four seconds of latency is considered excellent, while others find a quarter of a second of latency is okay and still others need less than a quarter of a frame. This impacts the ideal choice of codec, and devices can only implement so many of those.
Given the variety of requirements in the universe of AV-over-IP, standards like IPMX must be maximally generalized, while also remaining simple and effective. This is why IPMX is made up of the SMPTE ST 2110 standards, AMWA’s NMOS specifications, and a select group of technical recommendations authored by VSF. Taking into account everything that is specified within these documents, IPMX can cover enough use cases across enough industries that you really can start to see it as simply “video on the network”, no matter the application.
However, building equipment that would actually cover everything specified within IPMX would result in monolithic products of enormous expense. Alternatively, we can imagine a sea of products, each supporting various bits of IPMX, resulting in the promise of interoperability remaining frustratingly out of reach.
Real-world, multi-vendor interoperability means that an integrator, end user or system application can reliably determine if two products can connect to each other and exchange content, and that finding two matching products in the marketplace is a straightforward and logical endeavor. To “be IPMX,” a device must have a common set of things that are true about it, such as a common control protocol. But to deliver on the promise of affordable, interoperable products in the crazy mixed-up real world, grouping classes of IPMX devices into profiles is required.
In IPMX, the defining of profiles will include finding a balance between having enough choices to efficiently meet market requirements while avoiding too much fragmentation, which creates confusion and frustration. It also requires an awareness of the contexts and operational environments that drive the need for incompatible requirements to exist, and finding ways to isolate those requirements into logical buckets. Identifying how best to slice the market and what requirements MUST, SHOULD and MAY be met within each profile is both art and science.
IPMX will start with two profiles, both targeting the Pro AV market with compressed and uncompressed video that achieves sub-frame latency. An additional profile to cover WiFi / WAN and low-bandwidth applications is likely to follow soon after. Beyond this, there may be a need for additional profiles to cover use cases that don’t quite fit the technologies chosen for the initial offerings, or as new applications for AV-over-IP emerge.
While creating and marketing profiles carefully is important, IPMX will also provide what is needed for devices to negotiate connections, based on the capabilities of each device and network environment. These capabilities may go outside the bounds of IPMX, including proprietary codecs and control protocols. This flexibility supports the creativity of the marketplace and allows the standard to steadily evolve, independently of the hottest technology. This control plane feature is part of what has been accomplished within the AMWA IS-11 NMOS Stream Compatibility Management working group that is developing a specification to ensure applications and users can easily determine how to match receiver capabilities to what a sender can provide, including codecs, bandwidth, HDCP, resolution, and other properties.
Enabling seamless interoperability between devices made by different manufacturers in an ever expanding range of use cases and environments takes collaboration, time and expertise in multiple domains in multiple markets. If done correctly, an IPMX logo on a device, along with a profile designation, will bring confidence to people’s mind as they find the perfect solution to fit their project.