IPTV/VoD: The world that's on its way
Part one: Hitting the brakes before the crash
The trouble is that the internet is a garbled spaghetti mess of proprietary networks that are all set up in a completely different way. These networks meet at junctions (public or private peering points) to form the internet and conspire to deliver information in the quickest and most efficient way possible from one point to another. Some are multicast and QoS-enabled, some are not. There's no way to tell which route a piece of data will take when it makes it journey across the web of networks, and hence no way to guarantee its prompt delivery. There is very little standardisation between the operators of these networks, and efforts to unify them (such as the Mbone) have abjectly failed.
The irony is that most of these individual MPLS networks can do live television as it is. They already offer quality-guaranteed delivery of video from one place on their network to another. These point-to-point services are sold at a premium to private clients who want to get their TV signal from A to B reliably, for which the operators use multicast and common QoS systems, both of which are comparatively old technologies. Traffic running across these networks is generally "tagged" into two separate groups – public internet traffic, and private traffic. A private connection is typically implemented as a virtual private network (VPN) and can also use virtual local area network (VLAN) technology.
There is also more than enough capacity in transoceanic fiber to be to withstand the onslaught of live television and unicast video-on-demand. The fiber itself is already laid, but the cost comes through the capital expenditure for equipment to "light up" additional wavelength capacity at termination points. Scientific improvements in fiber-optic technology mean that this equipment gets more efficient as the years go on, that the demand strips its price and we can keep squeezing more and more bandwidth out of the same cable. Conservative anecdotal evidence suggests that less than six per cent of the available fiber is currently being used, and of that, only 60 per cent (~4 per cent of the total) is constantly at work. In third world countries, they are buying it for pennies by the metre and stringing it up in the air between individual houses.
To put this into perspective, the highest reported simultaneous traffic load passing through the LINX exchange in London's docklands is just over 100Gbit/s. To transmit all the visible channels on the Sky Digital platform in MPEG-4 AVC, you would need around 1Gbit/s (this of course excludes test channels, interactive applications, additional video tracks/camera angles and other hidden transmission variables). The first iteration of unicast video-on-demand being succeeded with new super-scalar multicast variants also means that if 1,000 people are all requesting the same movie title within the same 15 mins, less than 20 unique streams are needed to feed the rest of the audience parasitically. Unfortunately, the vested interests of the vendors lie with shipping as many units as possible, so efficiency doesn't do them many favours.
But unfortunately proprietary, money-grabbing, risk-aversion fever has gripped the carriers. They make more money selling private point-to-point video connections on their shiny MPLS networks and have very little interest in working with their competitors to interconnect live television services across the internet. Their primary goal is turning video carriage into a premium service that content providers and consumers pay through the nose for so they can appease investors and squeeze the last drop out of their past capital investments in infrastructure. For the next generation of the internet (and hence the next generation of profits), we need video to be a commodity.
On the surface it's a perfectly reasonable business decision. But look further down the road, and it's absurd. These companies are in business today because of co-operation, integration, and interoperation alone. If we'd have adopted the same policy at the beginnings of the internet it wouldn't exist. We'd all be trying to connect to JANET because of a similar campaign by all of the individuals to monopolise the network. And back then, investment in a new fangled communications technologies such as the internet was far riskier when the demand or "proof" was just blue sky conjecture.
Openness, neutrality and interoperation got them where they are today, and are the only things that will get them to point B. Resisting it is like turning over and ignoring everything that's come before and sabotaging your own salvation. Telecoms is a fast-moving and risky business, and they are making a catastrophic mistake doing nothing about it - as everyone is doing the same like lemmings running to a cliff edge. It's short-sighted madness.
Digital TX Limited is a London-based provider of technology and consultancy solutions for interactive digital television and broadband media. Alexander Cameron can be reached at [email protected]
As well as co-ordinating the birth of the IPTV Consortium (IPTVC), Alex is now offering a great value one-day workshop course on IPTV and Video On-Demand (VoD) specifically for web and media professionals. It can help you get up to speed on the latest technologies, content deals, operators and applications across the world, and offer immense value in identifying both new opportunities and threats for your business and personal career. If you would like more information, call Alex on 07986 373177 or email [email protected] Readers who quote The Register as their source will receive a 10 per cent discount on the course fees.
Sponsored: Hyper-scale data management