WO2002011364A2 - Envoi de donnees par flux - Google Patents

Envoi de donnees par flux Download PDF

Info

Publication number
WO2002011364A2
WO2002011364A2 PCT/IL2001/000716 IL0100716W WO0211364A2 WO 2002011364 A2 WO2002011364 A2 WO 2002011364A2 IL 0100716 W IL0100716 W IL 0100716W WO 0211364 A2 WO0211364 A2 WO 0211364A2
Authority
WO
WIPO (PCT)
Prior art keywords
receiver
data
transmission
block
packets
Prior art date
Application number
PCT/IL2001/000716
Other languages
English (en)
Other versions
WO2002011364A9 (fr
WO2002011364A3 (fr
Inventor
Doron Rajwan
Eyal Lubetzky
Joseph Yossi Azar
Original Assignee
Bandwiz, Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from IL13762400A external-priority patent/IL137624A0/xx
Priority claimed from IL13811400A external-priority patent/IL138114A0/xx
Priority claimed from IL14050400A external-priority patent/IL140504A0/xx
Priority claimed from PCT/IL2001/000106 external-priority patent/WO2001057667A1/fr
Application filed by Bandwiz, Inc. filed Critical Bandwiz, Inc.
Priority to AU2001280046A priority Critical patent/AU2001280046A1/en
Priority to US10/343,541 priority patent/US7327761B2/en
Publication of WO2002011364A2 publication Critical patent/WO2002011364A2/fr
Publication of WO2002011364A3 publication Critical patent/WO2002011364A3/fr
Publication of WO2002011364A9 publication Critical patent/WO2002011364A9/fr

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0078—Avoidance of errors by organising the transmitted data in a format specifically designed to deal with errors, e.g. location
    • H04L1/0083—Formatting with frames or packets; Protocol or part of protocol for error control
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00—Arrangements for detecting or preventing errors in the information received
    • H04L1/0001—Systems modifying transmission characteristics according to link quality, e.g. power backoff
    • H04L1/0006—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the transmission format
    • H04L1/0007—Systems modifying transmission characteristics according to link quality, e.g. power backoff by adapting the transmission format by modifying the frame length
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
    • G06F11/08—Error detection or correction by redundancy in data representation, e.g. by using checking codes
    • G06F11/10—Adding special bits or symbols to the coded information, e.g. parity check, casting out 9's or 11's

Definitions

  • the present invention is related to the transmission of stream information over a communication channel.
  • An advantage of streaming is that there is substantially no delay at a receiver before the media can be previewed, providing the media stream is sent in a manner synchronized to the viewing. If a user is required to wait for a media stream to repeat itself, for example in a carousel transmission system, the delay is as long as the transmission repeat time.
  • Another type of media dissemination method is multicasting, which may be combined with streaming.
  • a single copy of the media stream is broadcast to a plurality of receivers.
  • a complete file is transmitted by multicasting, without streaming, or is repeatedly transmitted.
  • Another type of data dissemination method is "on-demand" transmission.
  • the transmission of data is synchronized with a request by a user.
  • This type of dissemination is provided by unicasting (point to point connection).
  • a receiver can select a start time, out of a small number of available times, at each of which times a complete retransmission of the data is performed.
  • a particular application of "on-demand” dissemination is cable broadcasting of movies, i some implementations, a complete file may be transmitted "on-demand", for viewing when convenient. In such a file transmission system, if a media file size is FS and an available transmission bandwidth is TB, a minimally expected delay before the file can be played back is FS/TB.
  • An aspect of some embodiments of the invention relates to multicast transmission of a real-time stream of data, in which the data is comprised of a plurality of blocks transmitted as encoded or non encoded packets.
  • blocks that relate to currently available data are incomplete and are transmitted using unencoded packets or packets encoded differently from the first packets.
  • the different encoding is of each packet independent of the other packets.
  • a receiver is able to reconstruct a block using a mixture of encoded data and unencoded data.
  • when a block is completed its encoding method is changed, optionally, without any notification.
  • An aspect of some embodiments of the invention relates to an asynchronous media transmission method in which at least one packet is transmitted synchronously in order to speed up the transmission of preferential parts.
  • a receiver can immediately start continuous playback of the stream following reception of some synchronous packets.
  • a first block or blocks are synchronously transmitted for example by unicast in order to eliminate a delay which would be caused by the need to collect enough packets in order to reconstruct and/or decode the packets transmitted by an asynchronous transmission.
  • the rest of the blocks are optionally transmitted using a preferential multicast transmission method, for example, providing different effective transmission rates for different blocks.
  • a method of streaming data comprising: receiving an input stream of data, at a transmitter; dividing the input stream into a plurality of blocks, at least one of which blocks is incomplete; generating a plurality of first packets based on at least one block of data; generating at least one second packet from at least one as yet incomplete block of data; said block comprising recently received data ; transmitting said at least one first packet and at least one second packet to a receiver that can reconstruct said stream from said first packets and said second packets, said transmitting utilizing a differential protocol by which different parts of the data are transmitted at different rates, so that a receiver can join the transmission at any time and start receiving the data at a minimum delay; and generating at least one third packet from said at least one incomplete block; said at least one third packet being based at least in part on data received subsequent to data forming the basis for the at least one second packet; wherein said at least one second packet is transmitted at a higher rate than mandated by said protocol
  • the input stream is received by said receiver as it is being created.
  • the input stream is decodable by said receiver with a delay of less than the transmission and decoding time of one block, from said transmitter.
  • the receiver has a reception bandwidth less than 150%, 130%, 120% or 110% of a transmission bandwidth of said transmitter.
  • the at least one second packet is not FEC (forward error correction) encoded.
  • the at least one second packet is FEC (forward error correction) encoded to inter-relate data already sent for said incomplete block.
  • the at least one first packets are FEC (forward error correction) encoded.
  • the at least one first packets are sent using a plurality of carousels.
  • the at least one first packets are sent using a single carousel.
  • the at least one third packet is unicast.
  • the at least one first packets are multicast.
  • said transmission optionally comprises transmission over a TV broadcast system.
  • said transmission comprises transmission over a TV cable system.
  • said transmission comprises transmission over a TV satellite system.
  • said transmission is digital.
  • said transmission is analog.
  • said transmission comprises transmission over a computer network.
  • a receiver can join said transmission in an asynchronous manner and experience a delay of less than the transmission and decoding time of one block, relative to a designated entry point into said stream.
  • said entry point is a start of said stream.
  • the entry point is a current time point of said stream.
  • the method comprises reconstructing at said receiver, a block from a combination of said second packets and said third packets. In an embodiment of the invention, the method comprises, once the incomplete block is complete, transmitting the block in a same manner as the first packets only.
  • the method comprises reconstructing and playing back said stream by said receiver as a continuous stream.
  • said second packets and said at least one third packet are transmitted on different channels.
  • the receiver listens on fewer channels than said transmitter transmits on.
  • at least one channel is dedicated at said receiver for receiving said at least one third packet and comprising stopping receiving on said at least one dedicated channel once the incomplete block is completed at said transmitter.
  • the method includes starting receiving said second packets of said previously incomplete block on a different channel from said dedicated channel, at a time sufficient to prevent a delay in continuous playback by the receiver.
  • the method comprises notifying said receiver when said incomplete block is completed.
  • the method includes indicating in each packet its encoding.
  • the method includes dynamically distributing a plurality of reception channels of said receiver to receive packets from said transmission channels, in a manner that ensures a smooth playback of said input stream.
  • said second packets and said at least one third packet are transmitted on a same channel.
  • the method includes, as the incomplete block is completed, less information is send for those portions of the block that were previously transmitted, relative to the newer portions of the block.
  • a method of transmitting a data file over a communication medium comprising: determining desired relative reconstruction time frames for different parts of the file; allocating different transmission rates for the different parts of the file, responsive to said determining; dividing said file into sections utilizing a given protocol that inter-relates the relative reconstruction times and a transmission method; first transmitting at least a first section of said file responsive to a temporal synchronization of at least one receiver; and further transmitting the rest of the file, by multicasting, independent of the temporal synchronization of said receiver, to have effective transmission rates matching said different transmission rates, such that said parts can be reconstructed in their respective desired time frame.
  • said first transmitting and said further transmitting occur substantially together.
  • said first and said further transmitting utilize multiple channels.
  • said first and said further transmitting utilize a single channel.
  • said first and said further transmitting are interleaved utilizing said single channel.
  • said first section is unicast to said at least one receiver.
  • said receiver decodes said file at an average delay of less than the transmission and reconstruction time of said first section, if said first section were sent in a same manner as the other sections.
  • said receiver requests said first section.
  • said further transmitting comprises transmitting using at least one data carousel.
  • said further transmitting comprises transmitting using a forward error correction code.
  • a method of transmitting and receiving a data file in a manner substantially independent of a time of commencement of reception by a receiver comprising: transmitting a file, using a retransmission bandwidth, such that different parts of the file are retransmitted at different rates; and receiving said file using an effective reception bandwidth, smaller than said retransmission bandwidth, said transmitting being substantially independent of a time of commencement of said receiving; and commencing decoding to allow continuous playback of said file, at a delay from said retransmission, wherein said reception bandwidth is less than 80% of said retransmission bandwidth and wherein a ratio between said delay and a playback time of said file is smaller than 80%) of a ratio between said reception bandwidth and said retransmission bandwidth, wherein said receiver reconstructs said stream of data, in real-time, such that said receiver views the data of a block part of said file, prior to the entire block being transmitted.
  • the reception bandwidth is less than 50% or 20% of said retransmission bandwidth.
  • the reception bandwidth is less than 200%, 150% or 130% of a bandwidth required for smooth playback of said file using a synchronized transmission method.
  • said transmission is a multi-channel transmission and wherein said receiver is limited in a number of channels it can receive, said number being smaller than the number of channels being used for transmission.
  • receiving comprises optimizing said receiving, within the limits of said effective reception bandwidth, to minimize said delay.
  • receiving comprises limiting said reception to data required for current playback and for temporally limited sequential playback.
  • said data is reconstructed by said receiver as it is being created.
  • said data is at least partly decoded prior to the complete block being created.
  • said data is at least partly decoded with a delay of less than a transmission and reconstruction time of said block.
  • the method comprises unicasting a part of said block to said receiver.
  • at least one of said blocks is transmitted in an unencoded manner.
  • transmitting comprises multi-casting.
  • said encoding and transmission is such that an additional receiver can join the stream and experience a delay of less than the transmission and reconstruction time of a block relative to a designated entry point.
  • said transmitting comprises transmitting using FEC (forward error correction).
  • FEC forward error correction
  • said transmitting comprises transmitting using at least one data carousel.
  • Fig. 1 is a schematic illustration of a data streaming configuration, in accordance with an exemplary embodiment of the invention.
  • FIG. 2 is a schematic illustration of a file split up for transmission in accordance with an exemplary embodiment of the invention.
  • Fig. 3 is a flowchart of a method of reconstructing a transmitted data file, transmitted by streaming in accordance with an exemplary embodiment of the invention.
  • Fig. 1 is a schematic illustration of a data streaming configuration 100, in accordance with an exemplary embodiment of the invention.
  • One or more transmitters 102 transmit a file as a plurality of streams of data or data packets.
  • each such stream is multicast.
  • each stream is repeatedly transmitted, for example using a data carousel or a forward error correction code, as will be described in more detail below.
  • One or more receivers 104 receive the data streams and reconstruct a copy of the transmitted file.
  • An optional distribution controller 106 is described below.
  • a receiver 104 includes a local storage unit, for example one or more disks, for storing received parts of the data stream, until they are needed for reconstructing and/or display.
  • a local storage unit for example one or more disks
  • Fig. 2 is a schematic illustration of a file 200 split up for transmission in accordance with an exemplary embodiment of the invention.
  • file 200 is split into blocks 202, 204, 206, 208, 210 and 212, which are not all the same size.
  • each part of the file is larger by a factor than a previous part, for example a factor of two.
  • the last block may or may not fit this criterion, for example including only a residual portion of file 200.
  • the size of the factor may depend, inter alia, on the ratio between transmission speed and playback.
  • the factor can be lower than or higher than two.
  • the number of blocks into which file 200 is divided may be determined, for example, by the number of parallel streams available or the ability of the receivers to receive parallel streams. In some embodiments of the invention, as the number of blocks increases, the expected delay before the file can be played back is smaller. In some embodiments of the
  • the relationship between the number of blocks, BK, and delay time is
  • FILE RECONSTRUCTION Fig. 3 is a flowchart 300 of a method of reconstructing a transmitted data file, transmitted by streaming in accordance with an exemplary embodiment of the invention.
  • a plurality of N of the available K blocks into which the file is divided are received in parallel.
  • an I'th block is received, it is displayed (306), while continuing to receive the other blocks in parallel. Generally, as the blocks are in ascending order, the blocks will also complete reception in order. If a block is missing, some frames may be skipped, or the playback delayed until the required blocks are received.
  • the time that it takes to receive and display a block is the same as the time it takes to receive the next block.
  • the factor is smaller than 2.
  • channels are received at a rate lower than the bandwidth required for real-time playback.
  • the size of the smallest block and/or the complete block distribution are selected to achieve a desired expected delay.
  • the block size(s) are selected in conformance with transmission channel limitations and/or limitations on the availability of multicast address names.
  • the blocks are sent as consecutive bits, possibly arranged in packets. In each stream, the bits are repeatedly sent. However, if any bits are missed, a complete cycle must be waited. In addition, in a streaming mode, a minimum expected delay is the time to receive a complete cycle.
  • the data is sent using a FEC (forward error correction) code, in which a message of N bits can be reconstructed if any N bits (possibly plus a small overhead) are received.
  • FEC forward error correction
  • Data can begin to be usefully accumulated from the very first received bit.
  • any bits are lost, the following bits can replace them.
  • Exemplary FECs are described for example in Internet draft number draft- ietf-rmt-bb-fec-02 (November 17, 2000), the disclosure of which is incorporated herein by reference.
  • a data packet is generated by XORing together a plurality of data sections from the file (each data section is possibly the size of a channel block, while the division into blocks described above, can be unrelated).
  • the selection of data sections to use in each packet can depend, for example, on the location of the section relative to the start of the file.
  • the percentage of data sections used for a packet in a particular file section (described below) or file block is smaller than 50%, for example, being less than 5%, 10%, 20% or 30%.
  • a set of equations is solved, using the received packets as input.
  • a random number generation seed may be provided with each packet, to indicate which data sections of the original files take part in the packet.
  • the file is divided into sections, and separate packets are generated for each section. These sections may overlap the file blocks or may be considerably smaller. Possibly, packets from earlier sections, may be sent at a higher rate than packets from later sections and/or the section sizes may vary along the file.
  • cross-section packets are also provided, which packets combine data between different sections. Such packets are useful in that they allow to propagate the reconstruction of file section into another file section, even if some packets are missing from the other section.
  • such packets may assist in providing a limited look-ahead ability and/or compensate for missing packets.
  • the cross-section packets are limited to file sections that have similar ordinal numbers.
  • same size blocks may be used, with higher transmission rates of packets from the earlier blocks of the file.
  • preferential reception can be achieved by sending packets relating to earlier blocks more often than packets relating to later blocks.
  • sufficient packets to reconstruct a first block of the file will generally be received sooner than packets required for reconstructing a later part of the file.
  • the expected delay indicates the expected amount of time to accumulate sufficient packets.
  • the above block size factor is translated, in this embodiment, into a relative packet transmission rate.
  • a combined measure of packet transmission rate and relative block size can be used, to control the relative temporal availability (at the receiver) of different parts of the file.
  • preferential encoding schemes in which one part of the file takes part in more packets than other parts of the file, may also be used to ensure faster reception of those parts.
  • Preferential encoding can be applied to any part of the file, and/or any size units. It should be noted that a same cost in overhead of preferential encoding may be used to provide a small number of bits with a high additional preference or a large number of bits with a small additional preference. Possibly, different bits in the file are provided with different preference levels, possibly many levels used in a single file. In an exemplary embodiment of the invention, the preference level of consecutive bits is smoothly decreasing, so that the transmission load of the bits is smoothly decreasing.
  • preferential encoding is used in the transmission of a single block of the file, to ensure that earlier bits of the block are available for decoding sooner than later bits.
  • preferential encoding is used in conjunction with block based or packet based preferential transmission methods, to achieve more optimal usage of the variable bandwidth, possibly approaching the above "e” based theoretical limit. This method may be used even if the file is transmitted as a single block.
  • bit level preferential encoding Another potential advantage of bit level preferential encoding is that the preference level of a bit can be changed simply by using it less or more often in later packets (equations).
  • CONGESTION CONTROL One potential problem in a streaming system is that congestion can form at various _ j ⁇ xi x- ⁇ dress .- i i i _ ⁇ ii_ _ • j.' - - * J— _ - - — ⁇ - •
  • the receiving agent can disconnect itself from some of the channels to eliminate or reduce congestion.
  • some channels include data not found in other channels, for example, data for reconstructing a higher quality stream.
  • the content is distributed between the channels thus packets from all channels are required for reconstructing the complete data.
  • channels with data that is only required at a later time can be disconnected temporarily and reconnected at a later time without effecting the final result if reconnected in time.
  • This method may also be useful in multi-resolution streams, in which the highest resolution requires all the channels to be attended to. In case of congestion over a long period some channels can be disconnected and the resolution allowed to decrease while maintaining continuos playback.
  • an anticipated bandwidth is not available or an unexpected event reduces the available bandwidth. This might cause an unexpected delay at the receiver, in that the bit is not available on time for display.
  • bits that are not available can be dropped, for example, in some multi-resolution video streams.
  • the stream may be paused until the bit, or a sufficient string of bits is available.
  • the bit may be estimated, for example, based on previous or neighboring bits.
  • the delay is avoided.
  • the size factor of blocks (or other method of controlling block sizes) in the file is made smaller than would be possible using the available bandwidth, so that a next block can be available for decoding sooner than the completion of presentation of a previous block.
  • a small delay in block presentation does not affect its display on time.
  • the original file when congestion is anticipated and/or detected during the transmission of a file, the original file is re-divided into a different set of blocks (e.g., larger blocks and/or a different size factor), so that smooth playback, is maintained, albeit at a lower rate and/or greater delay.
  • bit level preference setting or resetting is used, in that previously transmitted bits are transmitted less often in the future, so that a desired preference level is achieved for those bits.
  • One type of delay is caused by the usage of an asynchronic transmission system, where the transmitter transmits blocks independently of the need of the receivers.
  • the transmitter cannot send information that matches the needs of multiple disparate receivers.
  • a typical receiver will experience a delay dependent on the minimal block size used by the transmitter, on the average, the time that it takes to receive and decode a block.
  • the reception of a block can be started independently of the time of joining to a channel.
  • delay is reduced by using synchronous transmission for the first block of data.
  • packets that comprise the first block are transmitted by unicasting so that the receiver can display or otherwise use the received packets as soon as they arrive.
  • the packets are unencoded or encoded differently (e.g., corresponding to a smaller block size or referring to data that was already sent or will be sent shortly) than the other blocks, to allow the receiver to quickly utilize the packets.
  • different block sizes are chosen in order to implement synchronous transmission more efficiently.
  • Other encoding, such as encryption, may also be applied, to all packets.
  • the bandwidth of channels used for such unicasting is traded off with the bandwidth used for multicasting, to achieve a desired tradeoff between overall transmission efficiency and initial delay.
  • This tradeoff may also depend on the number of receivers.
  • groups of receivers are grouped according to their start time and they each receive a semi-personalized multicasting of the first block, that matches the receivers needs to within a time resolution dictated by the number of groups used.
  • a receiver can request specific packets from the transmitter for this first block.
  • the specific packets relate to the particular entry point into the file selected by the receiver.
  • the receiver asks for retransmission of bad or missed packets.
  • an importance of a bit is monotonicly determined by its position in the stream.
  • the importance of a bit may depend on other factors as well. For example, in a stream of MPEG, in each frame, some bits are more important than others.
  • the encoding method is tailored to the data type that is being transmitted. For example, a block of the file is divided into sub-blocks having bits of different importance levels, with each sub-block being transmitted at a different rate. Possibly, the transmission level of a sub-block of important bits from a later block is higher than the transmission rate of a sub-block with low importance bits from a previous block.
  • the stream is reorganized prior to being encoded so that it conforms to the rule that earlier bits in the sfream are more important than later bits.
  • the stream may be reorganized back to its original form (possibly, less any bits that did not arrive on time).
  • the reorganizing before the decoding and after the decoding are performed by data-dependent pre- and post- processing units (e.g., software and/or hardware).
  • data-dependent pre- and post- processing units e.g., software and/or hardware.
  • MULTIPLE ENTRY POINTS The above method is especially useful for files that are viewed starting at their beginning.
  • the file may be treated as a plurality of sub files, each with its own starting point having its own expected delay.
  • Multiple entry points can also be provided by varying the packet fransmission rate over the file, for example providing greater transmission rates at the desired entry points.
  • other preferential encoding schemes can be used, for example using earlier blocks of the file in a greater percentage of the packets.
  • a user may use the previously received packets for a playback function.
  • the file structure is inverted in time, with earlier blocks being short and or transmitted more often, so that playback can be rapid.
  • packets received and relating to later blocks of the file may be used for a limited preview, for example of a small number of frames.
  • a receiver may be able to receive in parallel a plurality of storage streams and may require a memory buffer for each stream, to make disk access (e.g., for temporary storage of the stream) more efficient.
  • Such a device (or transmission channel) may be limited, for example, in total receiver bandwidth availability, disk size, memory, number of streams that can be listened to in parallel and/or number of streams that can be written or read to the disk in parallel.
  • the receiver receives as many streams (e.g., of the higher priority ones) as it can in parallel, and as one block is finished being received, disconnects that stream connects to the next one and starts receiving packets from it.
  • a receiver may be only able to receive M times real time playback speed. While the information divided to K blocks is transmitted at N times real time playback speed (M ⁇ N).
  • Such a receiver would connect to L streams where L • , thus resulting with a delay time
  • the other limitations can be translated into a limitation on the number of channels that can be effectively received at a time.
  • a channel is not connected to (or data received at a lower rate by packet dropping) for lack of processing power at the receiver to handle the overhead of the extra channel, while reconstructing previously received channels.
  • SIZE FACTOR In an exemplary embodiment of the invention, some overhead time is provided for each received block to be reconstructed, thus allowing a non-ideal CPU to be used and/or allowing for temporary bad network conditions delaying packet reception. Alternatively or additionally, the received blocks are decoded continuously, preventing CPU load peaks.
  • the relative size factor is made smaller than 2.
  • the relative size factor is made greater than 2.
  • real time playback with relatively short delays are achieved even though each streaming channel is the same speed or slower than the playback speed.
  • the factor then may be, for example, 1+N/K.
  • two or more of the streaming channels may be mixed into a single channel.
  • some fransmission channels may be faster than others (in practice).
  • the number of blocks, the size of the blocks and/or the relative size factors are dependent on the relative speed of the various channels.
  • the channel rates are monitored in real-time and the size of blocks modified accordingly, for example using distribution controller 106.
  • the method of differential transmission rates for different parts of the file may be usefully applied using single channel multicasting, in which packets relating to different parts of the file are selected for fransmission at a different relative rate.
  • the blocks do not overlap.
  • the blocks into which the file is divided there is at least some overlap between the blocks into which the file is divided.
  • at least some of the bits in the file are encoded to have a higher probability of being decoded sooner than the rest of the file. This can allow beginning playback of the next block even before it is all received.
  • overlap between file blocks is provided by at least some of the packets being cross-block packets. Such packets can assist in reconstructing the next block. Alternatively or additionally, such packets can be directed to the beginning of the next block.
  • cross-block packets use the solution of one channel to assist in solving equations for another channel.
  • a distribution controller 106 may be provided to decide which data files are streamed and/or multicast and/or what expected delay to offer. Such a controller may base its actions, for example, on request and/or responses from receivers 104 and/or channel limitations. Controller 106 may also be used to allow receivers 104 to respond to the received data, for example, emulating an interactive HTTP connection.
  • a 128 minute movie may be provided with an expected delay of no more than 0.5 minutes, by broadcasting the movie on 8 regular channels, as described above.
  • the expected delay is 16 minutes.
  • the parallel-received blocks may be stored, for example, using a TiVo (or other television transmission recorder).
  • Such broadcasting could also take advantage of methods known in the art for targeting only parts of the cable network.
  • the multicasting is received at network nodes and then re-broadcast as needed.
  • FEC coding may be used, in some embodiments of the invention, dropped frames may not be a problem and no coding is used.
  • the data is encrypted and/or compressed prior to transmission.
  • the communication network is a satellite, which typically has associated delay and link problems of sending a request for a specific media file.
  • the communication network is the Internet, where, for example, a movie server may desire to maintain constant data transmission rates, without being required to respond to requests by starting to send the same movie at multiple times. In some cases, when multiple requests arrive, additional channels are allocated to the movie, significantly reducing the expected delay time.
  • the communication network is a cellular telephone network or a radio network, where a user may desire to flip between channels, and always start at a beginning of a presentation item.
  • a packet based transmission network may be used, the above method can also be applied to other types of networks, including both synchronous and asynchronous networks and packet based, switching based and/or continuous transmission networks. Also, the above method may be applied to both digital and analog communications.
  • the transmitter transmits an index of the channels and their mapping into media presentations.
  • controller 106 may use responses to this index, to decide which files to broadcast, at what rate, how many and which entry points and/or how many sections to divide the file into.
  • a receiver e.g., a television, set top box or a computer
  • an exemplary channel may include some packets relating to other channels, for example channels of related content and/or language.
  • a feature of some types of broadcast channels is that their contents change only slowly over time. For example, news channels often continuously transmit a same content, while changing a small number of news items, every so often.
  • a differential decoding ability is used to selectively receive and decode only enough packets for displaying the changes in an item.
  • copies of previously received packets are stored, together with a code indicating the file version to which they apply. Only packets to the parts of the file updated in a newer version need to be downloaded and the old packets can be reused.
  • the file itself is used as a partial solution for recovering data from received packets, thus reducing the number of packets to be decoded.
  • This method can be used, for example, when a FEC is used, of the type where each packet is a XOR of a plurality of packets.
  • the version number of such a packet is the newest version number of any block used for the packet.
  • a set of equations linking together blocks and packets is solved. The old data may be used to assist such a solution.
  • Two particular examples of such a slowly changing channel is the Internet WWW page of CNN (which is widely viewed) and the CNN daily continuous newscast.
  • some changes may occur as result of the personalization of the channel to a particular user and/or as a result of a request made by the user.
  • sending the channel using the methods described above two advantages can be achieved. First, a short delay for retrieving most of the relevant channel is achieved, using a relatively low bandwidth. Thus, only the differences for particular viewers need to be sent. These differences can be sent, for example, by broadcast or by unicast (in Internet) or as data packets (in television) to be reconstructed by the receiver for the particular viewer.
  • differential decoding can be used to allow a receiver to receive only a small number of packets and use these packets to display the personalized/changed page.
  • a multicasting of WWW pages is used, together with a controller that receives responses from users, as a means for supporting an HTTP protocol using multicasting.
  • bits are overridden by providing new packets for the same bits, with other bit values.
  • the new bit values may be used in stead of the old ones.
  • the new bit vales may be added to the old bit values, generating a set of over- constrained equations. If the new bits are in the majority, the solution will be the new data.
  • the new bits are provided at a higher bit rate than other bits, for example, using preferential encoding techniques. Alternatively or additionally, whole packets including replacement bits are sent at a higher rate.
  • a message indicating that old cross-packet buckets or old packets are stale is provided as well.
  • a time period is defined, for example , in the data packets, that indicates that the packet is stale once the time passes, unless otherwise indicated.
  • a separate channel may be provided with such staleness indicators.
  • the new packets include, for example in the header, an indication that old packets are stale and/or that the new packet is a replacement packet.
  • the packet includes an indication of the bit selection method used for the new packet, especially if a different type of preferential encoding is used.
  • the above data streaming methods are used for streaming of real-time generated data.
  • the current viewing time is considered an entry point where minimum delay is desired,. This may be achieved for example, by using a minimum block size for the current data.
  • data that is being transmitted as it is generated is not encoded using a FEC code, while earlier, previously data is FEC-encoded. This allows the current data to be displayed at a shorter delay or no delay.
  • the beginning of the stream is also defined as an entry point. Possibly, additional entry points are defined.
  • the effective transmission rate of bits for the current time goes down until a generally desired bit rate based on the position in the file is achieved. This reduction may be smooth or it may be step-wise.
  • the user when a user joins a real-time event, the user receives data in parallel from multiple channels, possibly at a higher than the streaming data rate, possibly in order to catch-up with respect to acquiring the data sfream.
  • a user may activate fast forward and/or frame skipping functions (e.g., skip advertisements) in order that his display also catches up with the real-time event.
  • a user can select a different starting point in the stream.
  • the channels that relate to later parts of the event cannot contain data for their respective parts.
  • Such channels may be unused, maintained at a low data rate and/or used to assist in catching up, by transmitting unencoded information or provide additional bandwidth for transmitting previous information.
  • the transmission rate for all stream parts can be updated to reflect a desired reception and/or expected delay behavior, for example, using preferential encoding or transmission methods.
  • a particular receiver may be viewing a rerun of the event, after many other viewers have viewed it.
  • the packet transmission probabilities are adapted to take into account sections where entry may be desirable and/or sections where a playback function is desirable. These sections may be selected by an operator and/or automatically responsive to requests from viewers.
  • the above streaming methods are integrated with other data streaming methods.
  • codes used for streaming for example for compression, are used.
  • header sections of the sfream and/or periodic key frames are transmitted on a high- priority channel, to allow reconstructing the stream, from its middle.
  • a file to be transmitted is divided into blocks of non decreasing size. These blocks are transmitted optionally as FEC encoded packets on different channels.
  • a receiver can join the fransmission at any time and registers to receive in parallel a subset of the transmitted channels. When the receiver has enough packets to decode a particular block the receiver disconnects from one channel and connects to a new channel (with packets for a different block) instead.
  • the "current" block is incomplete, for example for reason of the data not existing, due to a delay in encoding or for any other reason.
  • such incomplete blocks are transmitted using unencoded packets, and once completed, encoded packets are used, optionally on a same channel.
  • a block can be reconstructed from a combination of encoded and unencoded blocks, using the unencoded blocks as solutions for the equations presented by the encoded blocks.
  • all or most packets are encoded, with the packets including reference only to parts of the block that exist. As the transmission proceeds, the frequency of such reference may go down, and the frequency of reference to later parts of the block increase.
  • the number of channels used for transmitting the unencoded data is such that the total transmission rate of these channels is equal to the stream creation rate or slightly higher.
  • channels are dedicated for a particular block.
  • some channel capacity may be used for packets from other blocks.
  • a receiver registers, for example, to the same number of channels as described above for the non-real-time case.
  • the first channels received by the receiver may be composed of completed blocks, which are optionally fully encoded.
  • the user can optionally catch up with the transmission of unencoded channels.
  • a block fills up while being transmitted as unencoded packets and the transmitter will encode the block and start transmitting encoded packets on the channel. Due to the additive nature of the coding scheme as explained above regarding differential encoding, receivers at different stages of reception can continue receiving without being delayed by the transition.
  • a receiver that has finished receiving all parts of the file prior to the current real time data can stay registered to the number of channels needed to keep up with the stream creation rate.
  • the contents of the stream are optionally unencoded packets.
  • the receiver will always have finished receiving a block before it needs to be displayed.
  • the receiver may be limited by the rate of creation of data and the packet size. In some cases, the data in the block will be displayed before the block is completed.
  • An exemplary choice of block sizes which will meet these requirements have the following properties: For a transmitter transmitting at N times realtime playback speed with K blocks and a receiver receiving at M times realtime playback speed where M ⁇ N and the user is registered to L of the blocks, R is the transmission rate for every channel and follow the
  • a further exemplary method of reception of a real time fransmission which does not use a further constraint on block sizes as the previous method is that the receiver registers to receive blocks as in a fully encoded transmission as described above. Free channels designated to receive non existent data are utilized to receive the real time (and possibly unencoded) data.
  • the current block of real time data is completed and encoded its reception is discontinued until the time when the amount of data received so far from this block is equal to the amount which would have been received if it had been fully encoded from the beginning of its reception.
  • all free channels receive data from the new current block of real time data.
  • a carousel or multiple carousels are used on multiple channels instead of forward error correction encoding.
  • the length of the real time event is not known ahead of time therefore the transmitter needs to update transmission parameters such as transmission speed and/or number of channels during the fransmission.
  • the viewer is given the option to start receiving from the beginning of the stream or from the position currently being created.
  • the viewer can simultaneously acquire blocks starting from the start of the sfream, so that the user can later perform a replay of the entire stream.
  • the viewer needs to receive at a higher bandwidth of at least the stream creation rate.
  • Exemplary applications of the system for real-time transmissions and non real time transmissions as mentioned previously include television broadcast systems or the Internet.
  • An exemplary implementation optionally uses various reception devices either harmless devices or smart devices such as a TV or a computer or other reception or recording devices identifying content for example by program ID.
  • these transmissions may be transmitted over cable networks or over satellite.
  • the system is implemented for transmissions of radio networks, cellular telephone networks, or standard telephone networks.
  • the system is implemented over a digital communication system.
  • the system is implemented over an analog communication system.
  • a plurality of transmitters may be provided, for example in different parts of the network, with each transmitter multicasting a different part of the file.
  • a geographical dispersion may reduce bottlenecks in the network.

Abstract

Procédé pour envoyer des données par flux qui consiste en ce qui suit: recevoir un flux de données par un émetteur; diviser le flux entrant en plusieurs blocs dont au moins un est incomplet; générer plusieurs premiers paquets basés sur au moins un bloc de données; générer au moins un deuxième paquet à partir d'au moins un bloc de données encore incomplet, ledit bloc contenant des données reçues récemment; transmettre au moins le premier paquet de données et au moins le deuxième paquet de données à destination d'un récepteur qui peut reconstruire ledit flux de données à partir desdits premiers et deuxièmes paquets, ladite transmission utilisant un protocole différentiel selon lequel des parties différentes des données sont transmises avec des débits différents, de manière à ce qu'un récepteur puisse rejoindre la transmission à tout moment et commencer à recevoir les données avec un délai minimal; et générer au moins un troisième paquet à partir dudit bloc incomplet. Ce troisième paquet est basé, au moins en partie, sur les données reçues après celles formant la base du deuxième paquet. Ce deuxième paquet est transmis avec un débit plus élevé que celui autorisé par ledit protocole, et ce afin de compenser la transmission répétée plus tard des informations dans ce premier paquet, avec un débit plus faible que celui autorisé par le protocole, une fois que le troisième paquet a été généré et transmis.
PCT/IL2001/000716 2000-02-03 2001-08-01 Envoi de donnees par flux WO2002011364A2 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
AU2001280046A AU2001280046A1 (en) 2000-08-01 2001-08-01 Data streaming
US10/343,541 US7327761B2 (en) 2000-02-03 2001-08-01 Data streaming

Applications Claiming Priority (13)

Application Number Priority Date Filing Date Title
IL13762400A IL137624A0 (en) 2000-08-01 2000-08-01 Media streaming
IL137624 2000-08-01
IL138114 2000-08-27
IL13811400A IL138114A0 (en) 2000-02-03 2000-08-27 Coding method
US24509800P 2000-11-02 2000-11-02
IL140504 2000-12-24
IL14050400A IL140504A0 (en) 2000-02-03 2000-12-24 Broadcast system
ILPCT/IL01/00106 2001-02-02
ILPCT/IL01/00105 2001-02-02
ILPCT/IL01/00107 2001-02-02
PCT/IL2001/000107 WO2001058131A2 (fr) 2000-02-03 2001-02-02 Systeme de diffusion
PCT/IL2001/000106 WO2001057667A1 (fr) 2000-02-03 2001-02-02 Transmission de donnees
PCT/IL2001/000105 WO2001058130A2 (fr) 2000-02-03 2001-02-02 Procede de codage

Publications (3)

Publication Number Publication Date
WO2002011364A2 true WO2002011364A2 (fr) 2002-02-07
WO2002011364A3 WO2002011364A3 (fr) 2002-05-16
WO2002011364A9 WO2002011364A9 (fr) 2003-12-11

Family

ID=27452524

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IL2001/000716 WO2002011364A2 (fr) 2000-02-03 2001-08-01 Envoi de donnees par flux

Country Status (1)

Country Link
WO (1) WO2002011364A2 (fr)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101889425B (zh) 2007-12-14 2013-10-30 汤姆逊许可公司 通过可变带宽信道进行同播的设备和方法
CN101889409A (zh) 2007-12-18 2010-11-17 汤姆逊许可公司 基于广播网络的文件大小估计设备和方法

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5379297A (en) * 1992-04-09 1995-01-03 Network Equipment Technologies, Inc. Concurrent multi-channel segmentation and reassembly processors for asynchronous transfer mode
EP0812110A2 (fr) * 1996-06-06 1997-12-10 International Business Machines Corporation Procédé pour la transmission séquentielle d'information vidéo à débit variable

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5379297A (en) * 1992-04-09 1995-01-03 Network Equipment Technologies, Inc. Concurrent multi-channel segmentation and reassembly processors for asynchronous transfer mode
EP0812110A2 (fr) * 1996-06-06 1997-12-10 International Business Machines Corporation Procédé pour la transmission séquentielle d'information vidéo à débit variable

Also Published As

Publication number Publication date
WO2002011364A9 (fr) 2003-12-11
WO2002011364A3 (fr) 2002-05-16

Similar Documents

Publication Publication Date Title
US7327761B2 (en) Data streaming
US20030007507A1 (en) Data streaming
WO2001057667A1 (fr) Transmission de donnees
US20200280595A1 (en) Apparatus, system, and method for multi-bitrate content streaming
US10681097B2 (en) Methods and systems for data transmission
TW560189B (en) Methods and apparatus for scheduling, serving, receiving media-on-demand for clients, servers arranged according to constraints on resources
US8301732B2 (en) Live media delivery over a packet-based computer network
US6725267B1 (en) Prefetched data in a digital broadcast system
JP2009506627A (ja) ダイナミックブロードキャストスケジューリングを使用したオンデマンドシステム及び方法
WO2000074367A9 (fr) Procede d'optimisation de video presque sur demande
US11870829B2 (en) Methods and systems for data transmission
JP2011530907A (ja) チャネル切替のためのメディアストリームの分離
AU2002322988B2 (en) System for delivering data over a network
WO2008134897A1 (fr) Procédé et système destinés à permettre une amélioration de la qualité du service dans des réseaux de transmission de données multimédia en continu
US20020138845A1 (en) Methods and systems for transmitting delayed access client generic data-on demand services
WO2002011364A2 (fr) Envoi de donnees par flux
WO2009026421A1 (fr) Appareil, système et procédé de transmission de données vidéo à l'aide d'un double flux multidiffusion, un flux étant retardé
JP2005506725A (ja) 遅延アクセスによるクライアントジェネリックなデータ・オン・デマンドサービスの伝送方法およびシステム
Poon et al. Interactive broadcasting system for VBR encoded videos
JP2012105341A (ja) ダイナミックブロードキャストスケジューリングを使用したオンデマンドシステム及び方法
Onifade et al. Guaranteed QoS for Selective Video Retransmission
EP1402331A2 (fr) Procedes et systemes d'emission de services sur demande de donnees generiques client a acces differe

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A2

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A2

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
AK Designated states

Kind code of ref document: A3

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A3

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG

REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

WWE Wipo information: entry into national phase

Ref document number: 10343541

Country of ref document: US

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 69(1) EPC

COP Corrected version of pamphlet

Free format text: PAGE 11, DESCRIPTION, REPLACED BY CORRECT PAGE 11

122 Ep: pct application non-entry in european phase
NENP Non-entry into the national phase in:

Ref country code: JP