Download: IP broadcasting

If you had to pick two letters from the English alphabet with the potential to have the greatest impact on the future of broadcasting, what would they be?

I would humbly suggest that the letters I and P are tightly coupled to almost every decision being made about the future of broadcasting.

IP (Internet Protocol) is used to refer to a group of emerging technologies that are reshaping the landscape of every industry involved with mass communications. Enabled by the IP concept, the Internet has emerged as a massive threat to TV, radio, cable, DBS and other broadcast distribution infrastructures.


TCP and IP were developed by a Department of Defense (DOD) research project to connect a number of different networks designed by different vendors into a network of networks (the “Internet”). It was initially successful because it delivered a few basic services that everyone needed (file transfer, electronic mail, remote logon) across a large number of client and server systems.

TCP/IP is composed of layers. IP is responsible for moving packets of data from node to node. IP forwards each packet based on a four-byte destination address (the IP number). The Internet authorities assign ranges of numbers to different organizations. The organizations assign groups of their numbers to departments. IP operates on gateway machines that move data from department to organization to region and then around the world.

TCP is responsible for verifying the correct delivery of data from client to server, as data can be lost in the intermediate network. TCP adds support to detect errors or lost data and to trigger retransmission until the data is correctly and completely received.

TCP/IP is not a broadcast technology. It is a one-to-one packet-based communications protocol designed for the reliable delivery of data across interconnected networks. Digital broadcasting, on the other hand, is a one-to-many stream-based technology designed for the isochronous delivery of data across a variety of competing, largely non-interconnected networks.

Isochronous means that the data packets within a stream must be delivered on time, and that the network must provide guaranteed bandwidth to support the peak bit-rate requirements of the content that is being delivered — typically audio and video streams. If the data does not arrive on time or it is corrupt, too bad — there are no second chances with real-time broadcast streams.

Given these realities, one might question why TCP/IP is so important to the future of broadcasting. After all, we have SMPTE 259M (SDI) to move digital video through the routing switchers found in modern video facilities. And the MPEG-2 transport protocol is optimized for the delivery of compressed digital video streams; it is the transport layer of choice for digital cable, DBS and DTV broadcasting around the world.

The answer is becoming obvious as the worlds of mass media broadcasting and the Internet collide. TCP/IP has become the language of peer-to-peer digital networking. It is found at the transport layer for the Ethernet networks that link computers together in offices and homes worldwide. And it is the transport layer for cable modems and digital subscriber lines, the broadband pipes that threaten to deliver mass media content on demand, to anyone, anywhere, anytime.

In a pre-NAB press conference, Sony demonstrated a new feature of their latest IMX studio digital video recorder — an Ethernet port that allows digital video streams to be delivered to all of those computer-based video tools that are popping up around digital broadcast facilities.

At a recent Society of Cable Engineers conference, keynote speaker Jim Chiddix, president of interactive personal video for AOL Time Warner, suggested that the cable industry needs to build one broadband network, probably IP-based, that can seamlessly unite all cable services on one network. A unified network simplifies network design and allows for improved bandwidth allocation among services.

The emerging reality is that IP networking is the driving force for the future of virtually every form of digital communications.

While traditional broadcasters are succumbing to the economic advantages of moving to an IP-based digital network infrastructure, the world of the Internet is succumbing to the one-to-many bandwidth-conservation advantages of broadcasting.

Logically, if you want to deliver a live streaming event over the Internet, it would be more efficient for a server to route one stream to many users. Unfortunately, most existing Internet routers cannot do this. Typically the server must duplicate the stream for every client.

It is not surprising that many engineers have been working on this problem, and that there are new protocols that will make broadcasting over the Internet a practical reality … eventually. At the top of the list is IP multicast.

IP multicast offers the ability to set up a variety of multicast services for any size group. Typically each user still initiates a transaction with the remote server. The server then adds their address to the routing information so that one stream of packets is routed to multiple destinations. Unfortunately, many routers do not support the IP multicast protocol, so IP multicast is only being used in subnetworks that support the protocol.

Today there are several ways to bypass the bottlenecks. One is a separate network known as the MBONE — in essence a virtual network that is layered on top of sections of the physical Internet — an interconnected set of Internet subnetworks and routers that support the delivery of IP multicast traffic.

Another approach is to mirror streaming media content on servers around the world, a concept pioneered by Cambridge-based Akamai. When you request a streaming media file from a Web site, you are connected to the closest mirror site. Even using TCP/IP, this can help to alleviate problems with congestion at key interconnection points on the Internet. If you are lucky enough to have a clear connection to the mirror site you may be able to “tune into” an IP multicast.

Today a large number of companies offer the ability to host IP broadcasts via the Internet. Real Networks has one of the most evolved business models, having locked up exclusive deals to carry live Internet radio broadcasts of major-league baseball and basketball, using a paid subscription model. One such company, — founded by Mark Cuban — was acquired by Yahoo, and has become Yahoo Broadcast.

Unrealized potential

The potential for IP broadcasting is difficult to measure, largely because of technical and competitive roadblocks erected by the entrenched industries that have risen to power using traditional broadcast technologies. The physical infrastructure of the Internet is a barrier that is likely to linger for years, until existing routers are upgraded to support the IP multicast protocol.

While there is a glut of wide area network capacity — witness the recently announced bankruptcy of Global Crossing — last-mile bandwidth continues to be a major issue.

Today broadband Internet services are largely controlled by the local cable and telephone monopolies. Provisions of the 1996 Telecommunications Act targeted at creating competition for broadband have proven to be ineffective. Broadband remains an expensive service — typically $40 per month and up — and the quality of service varies considerably. Intertainer, a web-based VOD movie service, has had to reject about half of its potential customers because their broadband service cannot reliably deliver the 500 kbits/s needed for the service.

Another issue is control of key patents for the compression of streaming media content. The owners of content and some key video compression patents have been somewhat reluctant to enable new competitors who want to use the Internet to bypass the traditional broadcast distribution infrastructures. Concerns about rampant piracy and battles over distribution rights and fees have been used to block and delay, even as the media giants develop their own Internet-based services.

Meanwhile, IP broadcasting is beginning to make inroads as traditional broadcast media upgrade to digital broadcast techniques based largely on MPEG-2 technologies. The cable industry is likely to migrate to IP broadcast techniques in the near future. And the major DBS systems, which are also offering satellite broadband services, may be a step ahead — DirecTV and DISH are both offering set-top boxes that take advantage of IP data delivery for new interactive services.

All of the groups behind the emerging standards for terrestrial DTV broadcasting — ATSC, DVB and ISDB — have been working on enabling standards for IP broadcasting. Work on the Multimedia Home Platform, the Interactive TV component of the DVB standard suite, is now addressing the delivery of IP data services. The A-90 data broadcast standard, developed by the ATSC, supports the delivery of IP data packets, and work is in progress on an IP multicast protocol optimized for terrestrial broadcast. And the Japanese ISDB system is being optimized to deliver IP broadcasts to a variety of wireless digital appliances.

Craig Birkmaier is a technology consultant at Pcube Labs, and hosts and moderates the Open DTV Forum.

Web links

Streaming media hosting services:

Akamai —
Digital Island —
Generic Media —
Real Networks —
Yahoo Broadcast —
Industry Directory of Streaming
Media —