IPTV: Creating a flexible network core
The practicalities of next generation services
"There's nothing decent on TV," is a common refrain in our house. Even though we have heeded the BBC adverts and invested in some Freeview boxes, it seems that even increasing the available channels from five to roughly 50 does not guarantee a cosy evening in front of the box. It wasn't much different when we lived in America, except that baseball was on most evenings; however, I am in a minority in my family being a fan of the sport.
There's obviously more exciting content available if you're ready to pay for it, either pay per view live events, or near video on demand films from subscription TV services such as Sky.
So I am always interested to hear people explain how they propose to make money in the emerging IPTV market. How will they deliver valuable content that goes beyond the same basic services as TV? If they don't, then it is difficult to see them achieving anything but driving down costs for packages of existing broadcast pay TV channels like HBO or UK Gold, leaving the high-value, satellite subscription services unaffected since Mr Murdoch has bought the rights to them.
One of the fundamental reasons for this is that both television and Near Video On Demand are multicast, or one to many, approaches to broadcasting: the media company decides what you want to watch and pushes it out. For alternative (i.e. not the traditional, state-owned) carriers this is the easiest type of service to offer because the same information is propagated across the network. In order to do this there is some engineering to accomplish, particularly setting up a fatter pipe into your house. That is why you see high speed internet being upgraded to 8 Megabits per second, using the latest DSL modem technology, providing more bandwidth for rich media files.
But today's digital media is moving in two different directions. The first is the delivery of High Definition content, which is much more demanding on network bandwidth, and the second is the rise of what might be called a social digital environment, including social networking sites, peer to peer voice video and data communications, and legal sharing of copyrighted content.
High definition digital content by itself can be broadcast in the same way regular TV is today. However, because it can only be watched by a limited number of viewers with the right equipment, it is something of a minority sport. The BBC's HD service is only available over satellite for example, because the bandwidth it would require over Freeview would use up too many channels and be watched by too few people, which would clearly be against their public service mandate.
Social digital networking represents a completely different challenge. It creates traffic flows that are much more like those found on a company network, with large files travelling both to and from individual endpoints. This idea on its own breaks the traditional concept of "oversubscription" in a consumer broadband deployment, where 20 homes may be sold 1 Megabit of downstream bandwidth, with only 2 Megabits of total bandwidth actually available which they all share (on the basis that not everyone will be downloading something at the same time).
It also runs into the "asymmetric" word in ADSL: home broadband has a much slower upload speed (as you may notice if sending a large file such as a photograph: it is slower to send than it is to receive). Clearly, in an environment where several people are watching one television stream this is fine. However, if everyone is watching something different, and possibly streaming out as well as in, the current network equipment and traffic models are no longer sufficient.
Recently, Juniper Networks, Intelliden, and their customer, Canadian carrier Telus, have been on the road talking about the good work they've been doing re-engineering the core of the carrier's network to counter such issues and create traffic flexibility. By combining the Juniper Networks Services Deployment System (SDX) with the Intelliden Dynamic Resource Provisioning solution, broadband services can now operate across the core of the network with a minimum of manual intervention and in a wide variety of traffic situations, reflecting exactly the type of patterns that are likely to arise with the evolution of the social digital environment.
The question remains, however, how carriers like Telus will develop and deliver services to the endpoints in their network that will allow them to offer more than a re-hash of the existing broadcast (and occasionally telephony services) currently available from cable companies. If they don't they run the risk of being marginalised, becoming merely the wholesale bandwidth provider ("bitpipe") to newer carriers who put together packages that combine ideas like cellular and home Wi-Fi telephony (termed "fixed mobile convergence") with subscription services to newer forms of digital content and social networking.
We can see from Juniper/Intelliden's example that the core networks can be given the necessary flexibility leveraging existing investments, but more work needs to be done on the development of compelling IP-based products and services like those outlined above to justify the further investment necessary at the very edge of the network.
If the carriers can get this right, they will be able to deliver such services universally to anyone with a phone line to their home.