“We run proof of concepts tests periodically with a broad range of vendors and for particular use cases and we'll do more in the year ahead,” explains Plunkett. “In some cases we're able to bring up a workable software-only implementation but it's far from a simple install and deploy model. It still takes quite a bit of tuning and collaboration with vendors.

“In many cases, vendors are taking product in fixed, known hardware configurations and deploying that in virtual machines only to find that the solution needs some serious software remedy to allow the product to operate in the much lesser known IT/IP environment,” he says.

“We have seen functionality and performance issues. The challenge is to create software that sits on generic IT infrastructure and still be as powerful as today's hardware-based product in order to make the switch economically attractive. We know that making software that can run on generic infrastructure is not easy so it's not a surprise to us that the industry is taking time to commit.”

On a positive note, Plunkett observes significant changes over the past year. “A year ago, getting anything to run in any fashion as software was a tall order of itself,” he says. “The pace of improvement is quite satisfactory. albeit that the marketing messages appear to have got out of step with the engineering reality.”

ITV recently extended its playout contract with Ericsson until 2024 and earlier this year the BBC signed its own seven-year extension with the company.

“Broadcasters are very keen to take advantage of the many possibilities that cloud opens up but most of those organisations will also be cautious, and maybe somewhat skeptical, about where we are right now,” he says.

“There is certainly not a blind appetite to move away from what is currently installed because the benefits of IP and software are not yet fully understood. IP and software defined networks should lead to flexibility of deployment, lower costs and the ability to change to different economic models but this is an area where the industry is finding its feet.”

Plunkett highlights two phases of evolution on the way to an IP future. “The version most of the industry is focussed on today emulates the current broadcast architecture but in on IP using software deployment. This brings definite benefits such as bringing channels to market much more quickly once you can template a channel to reduce cost.”

While significant, he says, this is not spectacularly game changing. “It does not fundamentally change the economics of the business. The evolution in the longer term is about really taking advantage of what IP and software can offer, rather than just porting across what we are familiar with today.”

He explains that a large proportion of the cost of running a channel is staffing “because of way we monitor channels and output and we have people ready to intervene in case of an incident or when schedule's change.” The current generation of IP and cloud product aren't really targeting those issues, he says. “Once we've got past the ability to deploy channels quickly there is an opportunity to take advantage of this capability.”

The channels most likely to adopt cloud playout first are non-live graphics-lite services.

“Some channels are less operationally demanding, less complex and less reactive and those are almost certainly better suited to the relatively immature new technology at this stage,” he says. “Once a particular type of channel is templates then the time it takes to initialise initialise a channel is less than 30 minutes. From own experiments, assuming all the work is done such as pre-integration and testing, then spinning up a new channel in seconds wouldn't be an exaggeration.

He stresses: “For too long the industry has been overly cautious of IP and IT. At Ericsson we tread a mid ground where we desire to take advantage of it but we want to do so with all the KPIs and SLAs that client's expect.”