articles
Ed's View - IPTV
One of the most exciting DTV transmission technologies emerging within the past few years is IPTV (Internet Protocol - Television). IPTV is a modification of the standard TCP/IP (Transmit Control Protocol/Internet Protocol) adapted to transmit "streaming" video. It may be considered the video version of VoIP (Voice over Internet Protocol). The reason for recent emphasis of IPTV is the desire by traditional telephone companies to compete with Cable for the delivery of video services. This desire is spurred by the increasing adoption of telecommunication reform legislation at both state and national levels, removing the competitive barriers to the delivery of video service to homes and businesses. In addition, the growing market for the delivery of video via Internet, enabled by 3G cellular technologies, and the fact that broadband Internet is available to over 50% of US households, has further spurred the deployment of IPTV. Further, the ability to deliver broadband information over standard telephone (POTS or "Plane Ole Telephone Service") lines minimizes the so-called "last mile" costs of delivering broadband services to the home.
So, just what is IPTV, and how does it fit in with the rest of the growing digital TV technologies? To understand IPTV it is first necessary, at least on a high level, to review the basic concept of TCP/IP: With TCP/IP digital information representing the actual data payload is encapsulated in individual frames or "packets" of variable length. A key component of each frame is a complex data tag called a "header." (See illustration below).
TCP Frame (not to scale)
The header contains a series of data fields representing sender and receiver addresses and timing information. This information is used by the network (e.g., the Internet) to appropriately route the data to assure that all packets arrive in the correct sequence and without error. Since the Internet, by design, can send individual packets to their destinations via different paths, timing and data errors are anticipated. If the IP error-detection system detects such errors, the data is ordered to be retransmitted until it is correctly received. As a result of this action and the variability of the network path lengths, the corrected packets can arrive at the receiver at different times, therefore, out of sequence. These sequence errors are detected at the receiver and corrected. Such IP data transmission and assured data reception schemes are sometimes referred to as "guaranteed accuracy" protocols. Guaranteed accurate data is obviously necessary for the proper delivery and display of text and graphic information.
However, because of this guaranteed accuracy characteristic, TCP is not suitable for "streaming" data (i.e., continuous audio and video data). Attempting to stream audio and video using TCP results in unacceptable delays and interruptions, i.e. very low "Quality of Service" or QOS. To ameliorate these problems and allow an acceptable QOS for streaming data, systems based on a highly modified version of TCP, called User Datagram Protocol (UDP), are employed. (See illustration below).
UDP Frame (not to scale)
The basic UDP, in turn, has been highly refined and modified over the past few years to deliver consumer acceptable broadcast quality television service, including HDTV.
Essentially, the UDP concept involves using a modified TCP header without the logic required to resend data for error correction, thereby converting the TCP protocol from a "guaranteed accuracy" to a "guaranteed time" delivery format. However, all of these aforementioned UDP based refinements to improve QOS have resulted in format incompatibilities. So, just because the service is titled "IPTV" does not at all guarantee that it will work with any other "IPTV" system.
Another factor affecting IP compatibility is outside the IP protocol itself. IP may be thought of as "middleware," independent of the physical means of transmission and format used for the application data payload transmitted. Obviously, if the receiver responds to a different physical format (e.g. telephone lines, Cable, wireless, etc.) than the transmitter, no connection can be made. Therefore, the transmitting and receiving physical format standards must be compatible.
Another consideration relative to IPTV compatibility, again outside the IP protocol itself, is the compression codec used. There are many different compression codecs being used for IPTV, such as 3GP (used for cell phones), H.263 (for videoconferencing), RealPlayer, MGEG 4 AVC, MPEG 2 and WMV (Windows Media Video). At this time the most popular formats for fixed IPTV services are MPEG 4 AVC and WMV. (MPEG4 AVC appears to be leading this race.) Therefore, IPTV has a very broad definition with no fixed standard. Not only must the exact sub-set of IP protocol be specified for a particular network, but also the physical transmission and codec specifications must be compatible.
Let's take a relatively high-level look at a general IPTV system as applied to a traditional point-multipoint broadcast model. The heart of an IPTV residential delivery system is a very high bandwidth "backbone" medium such as fiber optic cable or WiMax-type wireless systems. Each channel or service (TV channels, broadband Internet, and or Voice over IP (VoIP)) has its own IP address. If the customer is connected to the high bandwidth (fiber) backbone via his existing telephone line (POTS twisted-pair), he selects his channel via a set-top-box (STB). The STB is actually a modem that "maps" the selected channel numbers to an IP address. The IP address information is sent upstream on the telephone line to a router that is tied directly to the fiber backbone network. The router then couples the selected IP addressed digital stream to the customer's phone line. Therefore, the customers' do not actually receive the full bandwidth of the service. The customer essentially receivers only one channel that is remotely selected by the STP and router. In that manner, the expensive "last mile" broadband infrastructure is avoided.
In addition, the customer can receive all of his communications services from a single physical source and service provider. Also, he can retain his traditional POTS service over the same line and use it simultaneously with his broadband service. Wideband digital services using POTS lines come under the broad heading of Asymmetric Digital Subscriber Line, or DSL.
Normally, the DSL customer can receive sufficient bandwidth over his POTS line to support at least three MPEG 2 standard definition (SDTV) channels plus broadband Internet and VoIP. The majority of those systems appear to be upgrading their codecs to support MPEG 4 AVC in order to support HDTV.
Obviously, the advantages of applying this "switch IPTV" model to television service are very appealing. First, the customer not only automatically has the advantages of the inherently interactive capabilities of IP and access to the vast Internet, but also has the convenience economies of dealing with only one "telecommunications" provider.
As previously mentioned, DSL is not the only IPTV delivery format. An increasing number of providers are offering wireless services, using various forms of WiMax transmission protocols. In these systems the full backbone bandwidth is delivered to the home with the router at the receiver premises location. Also, an increasing number of homes have direct fiber optic line service. Further, delivery systems such as those using residential power lines are being deployed.
All of this competition to deliver IPTV and other broadband services to the homes and businesses are very good news for HDTV. No doubt, any IPTV system that does not have, or is not planning for HDTV upgrades, will not be competitive.
Ed
TCP Frame (not to scale)
The header contains a series of data fields representing sender and receiver addresses and timing information. This information is used by the network (e.g., the Internet) to appropriately route the data to assure that all packets arrive in the correct sequence and without error. Since the Internet, by design, can send individual packets to their destinations via different paths, timing and data errors are anticipated. If the IP error-detection system detects such errors, the data is ordered to be retransmitted until it is correctly received. As a result of this action and the variability of the network path lengths, the corrected packets can arrive at the receiver at different times, therefore, out of sequence. These sequence errors are detected at the receiver and corrected. Such IP data transmission and assured data reception schemes are sometimes referred to as "guaranteed accuracy" protocols. Guaranteed accurate data is obviously necessary for the proper delivery and display of text and graphic information.
However, because of this guaranteed accuracy characteristic, TCP is not suitable for "streaming" data (i.e., continuous audio and video data). Attempting to stream audio and video using TCP results in unacceptable delays and interruptions, i.e. very low "Quality of Service" or QOS. To ameliorate these problems and allow an acceptable QOS for streaming data, systems based on a highly modified version of TCP, called User Datagram Protocol (UDP), are employed. (See illustration below).
UDP Frame (not to scale)
The basic UDP, in turn, has been highly refined and modified over the past few years to deliver consumer acceptable broadcast quality television service, including HDTV.
Essentially, the UDP concept involves using a modified TCP header without the logic required to resend data for error correction, thereby converting the TCP protocol from a "guaranteed accuracy" to a "guaranteed time" delivery format. However, all of these aforementioned UDP based refinements to improve QOS have resulted in format incompatibilities. So, just because the service is titled "IPTV" does not at all guarantee that it will work with any other "IPTV" system.
Another factor affecting IP compatibility is outside the IP protocol itself. IP may be thought of as "middleware," independent of the physical means of transmission and format used for the application data payload transmitted. Obviously, if the receiver responds to a different physical format (e.g. telephone lines, Cable, wireless, etc.) than the transmitter, no connection can be made. Therefore, the transmitting and receiving physical format standards must be compatible.
Another consideration relative to IPTV compatibility, again outside the IP protocol itself, is the compression codec used. There are many different compression codecs being used for IPTV, such as 3GP (used for cell phones), H.263 (for videoconferencing), RealPlayer, MGEG 4 AVC, MPEG 2 and WMV (Windows Media Video). At this time the most popular formats for fixed IPTV services are MPEG 4 AVC and WMV. (MPEG4 AVC appears to be leading this race.) Therefore, IPTV has a very broad definition with no fixed standard. Not only must the exact sub-set of IP protocol be specified for a particular network, but also the physical transmission and codec specifications must be compatible.
Let's take a relatively high-level look at a general IPTV system as applied to a traditional point-multipoint broadcast model. The heart of an IPTV residential delivery system is a very high bandwidth "backbone" medium such as fiber optic cable or WiMax-type wireless systems. Each channel or service (TV channels, broadband Internet, and or Voice over IP (VoIP)) has its own IP address. If the customer is connected to the high bandwidth (fiber) backbone via his existing telephone line (POTS twisted-pair), he selects his channel via a set-top-box (STB). The STB is actually a modem that "maps" the selected channel numbers to an IP address. The IP address information is sent upstream on the telephone line to a router that is tied directly to the fiber backbone network. The router then couples the selected IP addressed digital stream to the customer's phone line. Therefore, the customers' do not actually receive the full bandwidth of the service. The customer essentially receivers only one channel that is remotely selected by the STP and router. In that manner, the expensive "last mile" broadband infrastructure is avoided.
In addition, the customer can receive all of his communications services from a single physical source and service provider. Also, he can retain his traditional POTS service over the same line and use it simultaneously with his broadband service. Wideband digital services using POTS lines come under the broad heading of Asymmetric Digital Subscriber Line, or DSL.
Normally, the DSL customer can receive sufficient bandwidth over his POTS line to support at least three MPEG 2 standard definition (SDTV) channels plus broadband Internet and VoIP. The majority of those systems appear to be upgrading their codecs to support MPEG 4 AVC in order to support HDTV.
Obviously, the advantages of applying this "switch IPTV" model to television service are very appealing. First, the customer not only automatically has the advantages of the inherently interactive capabilities of IP and access to the vast Internet, but also has the convenience economies of dealing with only one "telecommunications" provider.
As previously mentioned, DSL is not the only IPTV delivery format. An increasing number of providers are offering wireless services, using various forms of WiMax transmission protocols. In these systems the full backbone bandwidth is delivered to the home with the router at the receiver premises location. Also, an increasing number of homes have direct fiber optic line service. Further, delivery systems such as those using residential power lines are being deployed.
All of this competition to deliver IPTV and other broadband services to the homes and businesses are very good news for HDTV. No doubt, any IPTV system that does not have, or is not planning for HDTV upgrades, will not be competitive.
Ed