WO2011090918A2

WO2011090918A2 - Highly utilized communication channel with order and retransmissions

Info

Publication number: WO2011090918A2
Application number: PCT/US2011/021487
Authority: WO
Inventors: Eyran Lida; Nadav Banet; Ariel T. Sobelman
Original assignee: Valens Semiconductor Ltd.
Priority date: 2010-01-19
Filing date: 2011-01-18
Publication date: 2011-07-28
Also published as: WO2011090918A3

Abstract

A communication link comprising: a first transceiver configured to transmit a first set of packets at a predetermined rate with a first error resistance level, store the transmitted data in a buffer, receive a retransmission request, and retransmit the relevant data using one or more packets encoded with a second error resistance level that is higher than the first error resistance level. And a second transceiver configured to receive the first set of packets, detect an erroneous packet, request retransmission of the erroneous packet, and forward the data received in the packets according to its original order approximately after a fixed delay.

Description

HIGHLY UTILIZED COMMUNICATION CHANNEL WITH ORDER AND RETRANSMISSIONS

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of U.S. Provisional Patent Application No. 61/296,112, filed January 19, 2010.

TECHNICAL FIELD

Some of the disclosed embodiments relate to communication systems, and more particularly, to using a communication link to its maximum bandwidth.

BACKGROUND Various error detection and correction techniques exist for assisting with correct communication over a wired, an optical, and/or a wireless communication link. For example, Forward Error Correction (FEC) encodes information prior to transmission in such a way that any errors occurring during communication may be identified and corrected upon reception. Automatic Repeat Request (ARQ) retransmits information that is deemed to have been erroneously received or not received at all. ARQ may either retransmit only a portion of the information, transmit appropriate FEC information, or any combination thereof. An explicit request for retransmission may be transmitted by a receiver to the transmitter upon detecting a communication fault. However, both techniques are most effective when the quality of the link is quite predictable. Untuned FEC may result in an unnecessary waste of bandwidth, and misused ARQ may result in a delay in the data communication.

Some known methods for transmitting data over a wireless digital communication link between a transmitting station and a receiving station include the steps of transmitting a first packet using a modulation scheme; monitoring if the transmitted packet has been correctly received; and transmitting, using a lower modulation scheme, a second packet associated with the first packet if the monitoring did not indicate correct reception. However, these techniques do not utilize the communication link to its maximum capability, and/or do not maintain the order of the packets, and/or do not feature fixed latency.

BRIEF SUMMARY

In one embodiment, a communication link comprising: a first transceiver configured to transmit a first set of packets at a predetermined rate with a first error resistance level, store the transmitted data in a buffer, receive a retransmission request, and retransmit the relevant data using one or more packets encoded with a second error resistance level that is higher than the first error resistance level; and a second transceiver configured to receive the first set of packets, detect an erroneous packet, request retransmission of the erroneous packet, and forward the data received in the packets according to its original order approximately after a fixed delay. Optionally, most of the packets comprise packet headers featuring a higher error resistance level than the packet payloads. Optionally, the average throughput and maximum burst of the retransmissions are controlled. Optionally, the first transceiver controls the retransmissions. Optionally, the second transceiver controls the retransmissions. Optionally, the maximum amount of retransmitted packets is determined so as not to interfere with the proper transmission of the first set of packets. Optionally, the maximal throughput of the transmissions of the first set of packets and the retransmitted packets approaches the maximum capacity of the communication link. Optionally, the retransmitted packets are transmitted with a priority higher than the priority of the first set of packets, whereby the higher priority reduces the latency of the communication link. Optionally, the communication is over a wired medium. Optionally, the communication is over a wireless medium. Optionally, the communication is over an optical medium. Optionally, the communication link is coupled to an end device that does not support retransmission. Optionally, the received data is forwarded after an approximately fixed delay with extremely low variance. Optionally, the fixed delay is in the order of the round trip delay. Optionally, the communication link carries a first and a second stream; at least some of the packets of the first stream are to be protected by retransmissions, and the packets of the second stream are not to be protected by retransmissions. Optionally, a network comprising at least three of the above described communication links ensures an approximately fixed latency with extremely low latency variance. Optionally, the network operates well with end devices that do not support retransmission.

In one embodiment, a receiver comprising: a packet buffer configured to store received packets; a packet processing element coupled to the packet buffer and configured to detect erroneous packets and send a retransmission request upon detection of an erroneous packet, wherein the retransmitted packet features a higher error resistance than the initial packet; the packet buffer is further configured to store at least one additional packet before storing the retransmitted packet; and the packet processing element is further configured to maintain the original order of the packets by inserting the retransmitted packet to its proper location. Optionally, most of the packets comprise packet headers featuring a higher error resistance than the packet payloads. Optionally, most of the packet headers are encoded using a lower modulation scheme than the modulation scheme used to encode their corresponding packet payloads. Optionally, the erroneous packets include missing packets. Optionally, the missing packets are identified using a packet identification code that is assigned according to a predefined series of codes; whereby the packet processing element identifies the missing packets by comparing the packet identification codes of the received packets with the expected packet identification codes of the series. Optionally, the receiver is configured to operate at a throughput close to its maximum possible throughput. Optionally, some of the received packets are to be protected by retransmission requests and some of the received packets are not to be protected by retransmission requests. Optionally, the packets comprise a field used to mark whether the packet is to be protected by retransmission, and in most of the packets the field features a higher error resistance than the packet payload. Optionally, the receiver receives packets belonging to a first and second streams; at least some of the packets of the first stream are to be protected by retransmissions, and the packets of the second stream are not to be protected by retransmissions. Optionally, a system comprising the above described receiver is configured to transmit the packets to the receiver over a communication link; wherein most of the packets carry time- sensitive data, and the system ensures approximately fixed latency with extremely low latency variance. Optionally, a network comprising receivers and transmitters as described above ensures approximately fixed latency with extremely low latency variance. Optionally, the network operates well with end devices that do not support retransmission. In one embodiment, a network switch comprising multiple transceivers supporting packet retransmissions with a higher error resistance than the initial transmitted packet; the retransmissions are initiated by the receiver side; wherein a network comprising at least two links and at least one said switch maintains the original order of the packets and features a short and approximately fixed latency with extremely low latency variance. Optionally, most of the packets comprise packet headers featuring a higher error resistance than the packet payloads. Optionally, the average throughput and maximum burst of the retransmissions are controlled. Optionally, the network carries one or more main streams, and the maximum amount of retransmitted packets is determined so as not to interfere with the proper transmission of the main streams. Optionally, the retransmitted packets are transmitted with a priority higher than the priority of most of the initial transmitted packets, whereby the higher priority mechanism reduces the network's latency. Optionally, the network carries a first and a second stream; at least some of the packets of the first stream are to be protected by retransmissions, and the packets of the second stream are not to be protected by retransmissions. Optionally, the network is configured to operate at a throughput close to its maximum possible throughput.

In one embodiment, a method for operating a communication link close to its maximum capacity while maintaining the original order of transmitted data, comprising: transmitting at least one stream of data at a predetermined rate, the transmitted data featuring a first error resistance level; storing the received data in a buffer; requesting data retransmission upon detection of an error; retransmitting the data with a second error resistance level that is higher than the first error resistance level; and inserting the retransmitted data into the proper location in the buffer so as to maintain the original order of the data. Optionally, the method is operated over each hop of a network that maintains the original order of the data and features a short and approximately fixed latency with extremely low latency variance. Optionally, the data is transmitted using packets and most of the packets comprise packet headers featuring a higher error resistance than the packet payloads. Optionally, the method further comprising modulating most of the packet headers using a lower modulation than the modulation used to modulate the corresponding packet payloads. Optionally, the detected error includes a missing packet error, and further comprising identifying the missing packet using a packet identification code that is assigned according to a predefined series of codes. Optionally, the method further comprising transmitting another stream for which no retransmissions are requested. Optionally, the method further comprising marking which of the packets is to be protected by a retransmission and which is not. Optionally, the stream comprises time-sensitive data, and the method ensures approximately fixed latency with extremely low latency variance. Optionally, the method further comprising receiving the at least one stream of data at the predetermined rate with the first error resistance level from an end device that does not support retransmission.

Implementations of the disclosed embodiments involve performing or completing selected tasks or steps manually, semi-automatically, fully automatically, and/or a combination thereof. Moreover, depending upon actual instrumentation and/or equipment used for implementing the disclosed embodiments, several embodiments could be achieved by hardware, by software, by firmware, or a combination thereof. In particular, regarding hardware, different embodiments of the invention could be achieved by variations in the physical structure. Additionally, or alternatively, regarding software, selected functions of the invention could be performed by a data processor, such as a computing platform, executing software instructions or protocols using any suitable computer operating system. BRIEF DESCRIPTION OF THE DRAWINGS

The embodiments are herein described, by way of example only, with reference to the accompanying drawings. No attempt is made to show structural details of the embodiments in more detail than is necessary for a fundamental understanding of the embodiments. In the drawings:

FIG. 1A illustrates one embodiment of retransmitting a packet whose header and tail feature a higher error resistance than its payload;

FIG. IB is a flow diagram illustrating one method for operating a communication link close to its maximum capacity while maintaining the original order of transmitted data;

FIG. 2 illustrates a network that implements the embodiments over each hop; FIG. 3 illustrates one embodiment of source and sink devices supporting retransmissions and fixed delay for at least some of the streams; and

FIG. 4A to FIG. 4D illustrate different sections of a packet featuring different error resistances as a result of implementing different encoding schemes.

DETAILED DESCRIPTION

In the following description, numerous specific details are set forth. However, the embodiments of the invention may be practiced without some of these specific details. In other instances, well-known hardware, software, materials, structures and techniques have not been shown in detail in order not to obscure the understanding of this description. In this description, references to "one embodiment" or "an embodiment" mean that the feature being referred to may be included in at least one embodiment of the invention. Moreover, separate references to "one embodiment" in this description do not necessarily refer to the same embodiment. Illustrated embodiments are not mutually exclusive, unless so stated and except as will be readily apparent to those of ordinary skill in the art. Thus, the invention may include any variety of combinations and/or integrations of the embodiments described herein. Also herein, flow diagrams illustrate non-limiting embodiment examples of the methods, and block diagrams illustrate non-limiting embodiment examples of the devices. Some operations in the flow diagrams may be described with reference to the embodiments illustrated by the block diagrams. However, the methods of the flow diagrams could be performed by embodiments of the invention other than those discussed with reference to the block diagrams, and embodiments discussed with reference to the block diagrams could perform operations different from those discussed with reference to the flow diagrams. Moreover, although the flow diagrams may depict serial operations, certain embodiments could perform certain operations in parallel and/or in different orders from those depicted. Moreover, the use of repeated reference numerals and/or letters in the text and/or drawings is for the purpose of simplicity and clarity and does not in itself dictate a relationship between the various embodiments and/or configurations discussed. Furthermore, methods and mechanisms of the embodiments will sometimes be described in singular form for clarity. However, it should be noted that some embodiments may include multiple iterations of a method or multiple instantiations of a mechanism unless noted otherwise. For example, when a controller or an interface are disclosed in an embodiment, the scope of the embodiment is intended to also cover the use of multiple controllers or interfaces.

FIG. 1A illustrates one embodiment of retransmitting a packet whose header and tail feature a higher error resistance than its payload. Certain types of packets (#1, #2, #3, #4) of a first transmission are encoded to have a packet error rate much worse than the target packet error rate, while the retransmitted packets are encoded such that their packet error rate is approximately equal to or better than the target packet error rate. The source 402 sends packets to sink 404 at a predetermined rate. The source 402 stores the data transmitted through the packets for a period of time approximately equal to or longer than the round trip delay. The sink 404 receives the packets, checks the received packets for errors and/or missing packets, and requests retransmission of the erroneous packets (where erroneous packets also include the missing packets). Meanwhile, the source 402 continues to send packets at the predetermined rate, and sends the next packet(s) (in this example, packet #3), until the retransmission of the erroneous packets (in this example, packet #2). The payload of the retransmitted packet #2' is encoded using error resistance higher than the error resistance of packet #2, in order to ensure that no additional retransmissions will be required. After sending packet #2' to replace the erroneous packet #2, the source 402 sends the next unsent packet #4 according to the original order. On the receiving side, the sink 404 maintains the original order of the data by replacing the payload of the erroneous packet #2 in buffer 410 with the payload of the retransmitted packet #2'.

Different sections of the packet may feature different error resistances as a result of implementing different encoding schemes, as illustrated, for example, by FIG. 4A to FIG. 4D and as discussed in US patent application publication No. US20080187028, which is incorporated herein by reference. Alternatively, different sections of the packet may feature different error resistances as a result of implementing Forward Error Correction (FEC) techniques.

In one example, the communication link illustrated in FIG. 1A is used to transmit HDMI signals. In order to exploit the link to its maximum extent, the first set of packets delivers the HDMI data with a packet error rate of le-3 and CRC-32, and the retransmitted packets are delivered with a packet error rate of le-11, which is better than the packet error rate of le-9 required by the HDMI standard.

FIG. IB is a flow diagram illustrating one method for operating a communication link close to its maximum capacity while maintaining the original order of transmitted data. The method comprising the following steps: In step 440 transmitting at least one stream of data at a predetermined rate; the transmitted data featuring a first error resistance level. In step 441, storing the received data in a buffer. In step 442, requesting data retransmission upon detection of an error. In step 443, retransmitting the data with a second error resistance level that is higher than the first error resistance level. And in step 444, inserting the retransmitted data into the proper location in the buffer so as to maintain the original order of the data. In one embodiment, the method is operated over each hop of a network that maintains the original order of the data and features a short and approximately fixed latency with extremely low latency variance. In one embodiment, the detected error includes a missing packet error, and further comprising identifying the missing packet using a packet identification code that is assigned according to a predefined series of codes. In one embodiment, the stream comprises time-sensitive data, and the method ensures approximately fixed latency with extremely low latency variance. FIG. 2 illustrates a network that implements the method illustrated by FIG. 1A over each hop. Thus, the entire network maintains the order of the transmitted packets, and requests for retransmissions are responded to using packets with improved error resistance and, for the most part, without adversely affecting the original streams transmitted from the sources 502 and 504. Optionally, the sources 502 and 504 transmit the original streams approximately according to a fixed transmission rate. Optionally, the network approximately ensures a fixed latency with extremely low latency variance. Optionally, the sources 502 and 504 receive the data from end devices that do not support retransmission.

In one embodiment, the switches 506, 508, and 510 of the network illustrated in

FIG. 2 comprise buffers and therefore the network can operate well with legacy end devices 502, 504, 512, and 514 that do not support retransmission. In this embodiment the total latency of the network may also be reduced when retransmission is not required over at least one of the hops.

FIG. 3 illustrates one embodiment of source and sink devices supporting retransmissions and fixed delay for at least some of the streams. On the source side, the source's scheduler 602 receives a first stream to be protected by the retransmission mechanism (marked by a solid line) and a second stream that is not to be protected by the retransmission mechanism (marked by a dashed line). The source's retransmission module 604 optionally adds additional packet identification data to the first stream packets in order to identify their sequential order, and then stores a section of the first stream in the source's buffer 606. The first and second streams are transmitted by the source's physical layer component 608 over the link 612, and are received on the sink side by the receiver's physical layer component 620. The receiver's retransmission module 622 forwards the packets of the second stream to the receiver's scheduler 626, with or without a CRC check, and forwards the packets of the first stream to the receiver's buffer 624. Upon detection of a CRC error in a first stream packet, or detection of a missing packet, optionally based on a gap in the additional packet identification data, the receiver's retransmission module 622 sends a proper retransmission request to the source. The source's retransmission module 604 extracts the required packet from the source's buffer 606, re-encodes the packet to feature a better error resistance, and sends the retransmitted packet to the sink. In order to maintain the original order of the packets, the receiver uses the receiver's buffer 624 to introduce latency on all protected packets, proper or erroneous, and the receiver's retransmission module 622 inserts the retransmitted packet into its original position in the receiver's buffer 624. After the required latency period passes, the packets from the receiver's buffer 624 are forwarded to the receiver's scheduler 626.

In one embodiment, the transceiver initially transmits some of the packets that should have been transmitted with a low error resistance level with higher error resistance level in order to prevent potential retransmissions.

In one embodiment, a transceiver on a first side of the link transmits a second set of packets that is not limited to be received in its original order. In this case, the transceiver encodes the second set of packets with a first low error resistance level, then the second transceiver on the other side of the link receives the second set of packets, detects an erroneous second packet, requests retransmission of the erroneous second packet, and forward the data received in the second packets according to the order it was received and not according to its original order.

In one embodiment, the communication link illustrated in FIG. 3 carries multimedia signals. The video pixel data is protected by retransmissions, while latency sensitive data that does not have to be synchronized to the video pixel data is transmitted over the second stream that is not to be protected by the retransmission mechanism, and therefore without the fixed delay. In this case, neither the transmitter, nor the receiver needs to store the unprotected packets in the retransmission buffer. In one embodiment, the retransmission protection mechanism is not required because the unprotected data types are transmitted using a much higher error resistance level, such as lower PAM, than the initial protected packets. For example, link control messages and video DDC signals and may be transmitted with a high error resistance and without the retransmission mechanism.

In one embodiment, a first video stream is protected by the retransmission mechanism and a second video stream is not protected by the retransmission mechanism. Therefore, the unprotected video stream features a shorter link delay than the protected video stream. In one embodiment, the maximal link delay determines the maximum number of retransmissions, which determines the amount of traffic to be protected by the retransmission mechanism.

There are cases where it is necessary to utilize a communication channel approximately to its maximum capacity, while maintaining the order of the transmitted packets and a fixed and short latency. In order to utilize the communication channel approximately to its maximum capability, the data cannot be protected only by Forward Error Correction (FEC) code, which consumes extra bandwidth when the interference over the communication link cannot be accurately predicted. Therefore, erroneous packets that cannot be corrected by the optional FEC code have to be retransmitted. And in order to maintain the order of the transmitted packets and a fixed and short latency, the retransmitted packets should be requested as soon as the receiver identifies an error in a received packet, and/or identifies a missing packet.

In one embodiment, a communication link comprising a first and second transceivers. The first transceiver transmits a first set of packets at a predetermined rate with a first error resistance level and stores the transmitted data in a buffer. The second transceiver receives the first set of packets, checks the packets for errors, and upon detecting an erroneous packet, requesting retransmission of the erroneous packet. The first transceiver receives the retransmission request and retransmits the relevant data using one or more packets encoded with a second error resistance level that is higher than the first error resistance level. Then the second transceiver forwards the data received in the packets according to its original order approximately after a fixed delay. In one embodiment, the average throughput and maximum burst of the retransmissions are controlled such as not to over a predefined amount, not to overload the link, not to interfere with the proper transmission of the first set of packets, To bring the maximal throughput of the transmissions of the first set of packets and the retransmitted packets to the maximum capacity of the communication link, and/or not to increase the delay above a certain value. The retransmissions may be controlled by first transceiver, the second transceiver, and/or any other appropriate element. In one embodiment, the retransmitted packets are transmitted with a priority higher than the priority of the first set of packets. Retransmitting the packets with a higher priority may reduce the latency of the communication link by shortening the time between requesting a retransmission and receiving the retransmitted packet. The embodiments may be implemented over a variety of mediums, such as a wired medium, a wireless medium, or an optical medium. Moreover, the embodiments may be used to create a network that ensures an approximately fixed latency with extremely low latency variance. Such a network may also operate well with end devices that do not support retransmission.

In one embodiment, a receiver comprising a packet buffer coupled to a packet processing element. The packet buffer stores the received packets and the packet processing element detects erroneous packets and sends a retransmission request upon detection of an erroneous packet. Optionally, the packets include a packet identification code that is assigned according to a predefined series of codes and the packet processing element identifies the missing packets by comparing the packet identification codes of the received packets with the expected packet identification codes of the series. When receiving the retransmitted packet, the packet processing element maintains the original order of the packets by inserting the retransmitted packet to its proper location.

In one embodiment, a network switch includes multiple transceivers supporting packet retransmissions with a higher error resistance than the initial transmitted packet. The retransmissions are initiated by the receiver side and a network comprising at least two links and at least one said switch maintains the original order of the packets and features a short and approximately fixed latency with extremely low latency variance. In one embodiment, the network carries one or more main streams, and the maximum amount of retransmitted packets is determined so as not to interfere with the proper transmission of the main streams. In one embodiment, the network's latency is reduced by retransmitting the packets with a priority higher than the priority of most of the initial transmitted packets.

Certain features of the embodiments, which may have been, for clarity, described in the context of separate embodiments, may also be provided in various combinations in a single embodiment. Conversely, various features of the embodiments, which may have been, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable sub-combination. The embodiments are not limited in their applications to the details of the order or sequence of steps of operation of methods, or to details of implementation of devices, set in the description, drawings, or examples.

While the methods disclosed herein have been described and shown with reference to particular steps performed in a particular order, it will be understood that these steps may be combined, sub-divided, or reordered to form an equivalent method without departing from the teachings of the embodiments. Accordingly, unless specifically indicated herein, the order and grouping of the steps is not a limitation of the embodiments.

While the embodiments have been described in conjunction with specific examples thereof, it is to be understood that they have been presented by way of example, and not limitation. Moreover, it is evident that many alternatives, modifications and variations will be apparent to those skilled in the art. Accordingly, it is intended to embrace all such alternatives, modifications and variations that fall within the spirit and scope of the appended claims and their equivalents.

Claims

WHAT IS CLAIMED IS:

1. A communication link comprising:

a first transceiver configured to transmit a first set of packets at a predetermined rate with a first error resistance level, store the transmitted data in a buffer, receive a retransmission request, and retransmit the relevant data using one or more packets encoded with a second error resistance level that is higher than the first error resistance level; and

a second transceiver configured to receive the first set of packets, detect an erroneous packet, request retransmission of the erroneous packet, and forward the data received in the packets according to its original order approximately after a fixed delay.

2. The communication link of claim 1, wherein most of the packets comprise packet headers featuring a higher error resistance level than the packet payloads.

3. The communication link of claim 2, wherein the average throughput and maximum burst of the retransmissions are controlled.

4. The communication link of claim 3, wherein the first transceiver controls the retransmissions.

5. The communication link of claim 3, wherein the second transceiver controls the retransmissions.

6. The communication link of claim 3, wherein the maximum amount of retransmitted packets is determined so as not to interfere with the proper transmission of the first set of packets.

7. The communication link of claim 3, wherein the maximal throughput of the transmissions of the first set of packets and the retransmitted packets approaches the maximum capacity of the communication link.

8. The communication link of claim 3, wherein the retransmitted packets are transmitted with a priority higher than the priority of the first set of packets, whereby the higher priority reduces the latency of the communication link.

9. The communication link of claim 2, wherein the communication is over a wired medium.

10. The communication link of claim 2, wherein the communication is over a wireless medium.

11. The communication link of claim 2, wherein the communication is over an optical medium.

12. The communication link of claim 2, wherein the communication link is coupled to an end device that does not support retransmission.

13. The communication link of claim 1, wherein the received data is forwarded after an approximately fixed delay with extremely low variance.

14. The communication link of claim 13, wherein the fixed delay is in the order of the round trip delay.

15. The communication link of claim 1, wherein the communication link carries a first and a second stream; at least some of the packets of the first stream are to be protected by retransmissions, and the packets of the second stream are not to be protected by retransmissions.

16. The communication link of claim 1, wherein the first transceiver is further configured to initially transmit some of the first set of packets with the second error resistance level in order to prevent a potential retransmission.

17. The communication link of claim 1, wherein the first transceiver is further configured to transmit a second set of packets encoded with the first error resistance level; and the second transceiver is further configured to receive the second set of packets, detect an erroneous second packet, request retransmission of the erroneous second packet, and forward the data received in the second packets according to the order it was received and not according to its original order.

18. A network comprising at least three communication links according to claim 1, wherein the network ensures an approximately fixed latency with extremely low latency variance.

19. The network of claim 18, wherein the network operates well with end devices that do not support retransmission.

20. A receiver comprising:

a packet buffer configured to store received packets;

a packet processing element coupled to the packet buffer and configured to detect erroneous packets and send a retransmission request upon detection of an erroneous packet, wherein the retransmitted packet features a higher error resistance than the initial packet;

the packet buffer is further configured to store at least one additional packet before storing the retransmitted packet; and

the packet processing element is further configured to maintain the original order of the packets by inserting the retransmitted packet to its proper location.

21. The receiver of claim 20, wherein most of the packets comprise packet headers featuring a higher error resistance than the packet payloads.

22. The receiver of claim 21, wherein most of the packet headers are encoded using a lower modulation scheme than the modulation scheme used to encode their corresponding packet payloads.

23. The receiver of claim 21, wherein erroneous packets include missing packets.

24. The receiver of claim 23, wherein the missing packets are identified using a packet identification code that is assigned according to a predefined series of codes; whereby the packet processing element identifies the missing packets by comparing the packet identification codes of the received packets with the expected packet identification codes of the series.

25. The receiver of claim 21, wherein the receiver is configured to operate at a throughput close to its maximum possible throughput.

26. The receiver of claim 20, wherein some of the received packets are to be protected by retransmission requests and some of the received packets are not to be protected by retransmission requests.

27. The receiver of claim 26, wherein the packets comprise a field used to mark whether the packet is to be protected by retransmission, and in most of the packets the field features a higher error resistance than the packet payload.

28. The receiver of claim 20, wherein the receiver receives packets belonging to a first and second streams; at least some of the packets of the first stream are to be protected by retransmissions, and the packets of the second stream are not to be protected by retransmissions.

29. A system comprising the receiver of claim 20, the system further comprising a transmitter configured to transmit the packets to the receiver over a communication link; wherein most of the packets carry time-sensitive data, and the system ensures approximately fixed latency with extremely low latency variance.

30. A network comprising receivers and transmitters according to claim 29, wherein the network ensures approximately fixed latency with extremely low latency variance.

31. The network of claim 30, wherein the network operates well with end devices that do not support retransmission.

32. The receiver of claim 20, wherein the packet buffer further configured to store a second set of packets; the packet processing element further configured to detect erroneous packets in the second set and send a retransmission request upon detection of an erroneous packet; and the packet processing element is further configured to store the retransmitted second set packet according to the order it was received and not according to its original order.

33. A network switch comprising multiple transceivers supporting packet retransmissions with a higher error resistance than the initial transmitted packet; the retransmissions are initiated by the receiver side; wherein a network comprising at least two links and at least one said switch maintains the original order of the packets and features a short and approximately fixed latency with extremely low latency variance.

34. The network switch of claim 33, wherein most of the packets comprise packet headers featuring a higher error resistance than the packet payloads.

35. The network switch of claim 33, wherein the average throughput and maximum burst of the retransmissions are controlled.

36. The network switch of claim 35, wherein the network carries one or more main streams, and the maximum amount of retransmitted packets is determined so as not to interfere with the proper transmission of the main streams.

37. The network switch of claim 35, wherein the retransmitted packets are transmitted with a priority higher than the priority of most of the initial transmitted packets, whereby the higher priority mechanism reduces the network's latency.

38. The network switch of claim 33, wherein the network carries a first and a second stream; at least some of the packets of the first stream are to be protected by retransmissions, and the packets of the second stream are not to be protected by retransmissions.

39. The network switch of claim 33, wherein the network is configured to operate at a throughput close to its maximum possible throughput.

40. A method for operating a communication link close to its maximum capacity while maintaining the original order of transmitted data, comprising: transmitting at least one stream of data at a predetermined rate, the transmitted data featuring a first error resistance level; storing the received data in a buffer; requesting data retransmission upon detection of an error; retransmitting the data with a second error resistance level that is higher than the first error resistance level; and inserting the retransmitted data into the proper location in the buffer so as to maintain the original order of the data.

41. The method of claim 40, wherein the method is operated over each hop of a network that maintains the original order of the data and features a short and approximately fixed latency with extremely low latency variance.

42. The method of claim 40, wherein the data is transmitted using packets and most of the packets comprise packet headers featuring a higher error resistance than the packet payloads.

43. The method of claim 42, further comprising modulating most of the packet headers using a lower modulation than the modulation used to modulate the corresponding packet payloads.

44. The method of claim 42, wherein the detected error includes a missing packet error, and further comprising identifying the missing packet using a packet identification code that is assigned according to a predefined series of codes.

45. The method of claim 42, further comprising transmitting another stream for which no retransmissions are requested.

46. The method of claim 42, further comprising marking which of the packets is to be protected by a retransmission and which is not.

47. The method of claim 42, wherein the stream comprises time-sensitive data, and the method ensures approximately fixed latency with extremely low latency variance.

48. The method of claim 42, further comprising receiving the at least one stream of data at the predetermined rate with the first error resistance level from an end device that does not support retransmission..