ADAPTIVER MULTIPFAD-SCHEDULER

Description

Technical field

[0001] The present invention relates to techniques for adaptively selecting a scheduling scheme for data packet transmission in a multipath communication system. Related aspects include a computer program, an adaptive multipath scheduler, and a multipath communication system. background

[0002] Methods such as Multipath TCP (Multipath Transmission Control Protocol, or simply Multipath TCP), Multipath QUIC (Multipath Quick UDP Internet Connections, or simply Multipath QUIC), or dual connectivity at the Packet Data Convergence Protocol layer in 3GPP (3rd Generation Partnership Project) are examples of well-known protocols that allow multiple transmission paths of a multipath communication system to be used for data packet transmission between two devices (between a sender and a receiver). However, some existing state-of-the-art methods use a single scheduling scheme during data packet transmission across multiple transmission paths.Furthermore, some state-of-the-art methods provide for the ability to enable or disable packet duplication during data transmission (the so-called packet duplication scheme) under certain circumstances. Depending on the current data transmission scenario (e.g., network load, signal strength of a sender, or a combination thereof), one or more scheduling schemes may achieve improved quality of service (QoS) compared to other scheduling schemes (e.g., in terms of transmission latency, reliability, security, required bandwidth, or a combination thereof). Some existing methods do not take this into account, as they only use one scheduling scheme.

[0003] EP 3 860 064 A1 discloses a method for selecting a scheduling scheme, wherein the scheduling scheme is selected based on the correlation of latency measurements of the communication paths.

[0004] Therefore, there is a need to develop new techniques for multipath communication systems that can solve some or all of the problems mentioned above. Summary of the invention

[0005] A first general aspect of the present disclosure relates to a method for adaptively selecting a scheduling scheme for data packet transmission in a multipath communication system. The method is defined in claim 1.

[0006] A second general aspect of the present disclosure relates to a computer program designed to perform the procedure according to the first general aspect of the present disclosure.

[0007] A third general aspect of the present disclosure relates to an adaptive multipath scheduler of a multipath communication system designed to execute the method according to the first general aspect of the present disclosure and / or to execute the computer program according to the second general aspect.

[0008] A fourth general aspect of the present disclosure relates to a multipath communication system comprising an adaptive multipath scheduler according to the third aspect. Furthermore, the multipath communication system according to the fourth aspect comprises a transmitter designed to transmit a plurality of data packets to the receiver. The transmitter of the fourth aspect can be connected to the receiver via two or more transmission paths. The multipath communication system of the fourth aspect is designed to perform the method according to the first general aspect of the present disclosure.

[0009] The techniques described in the first to fourth general aspects may have one or more of the following advantages.

[0010] First, the techniques presented here allow for the adaptive selection of a scheduling scheme for data packet transmission in a multipath communication system, compared to other existing scheduling schemes, by determining the properties of the various transmission paths of the multipath communication system (such as different transmission path parameters and possible correlations between transmission path parameters of different transmission paths) in real time. As a result, the quality of service for data packet transmission over multiple transmission paths can be better than with some other state-of-the-art methods.

[0011] Secondly, the techniques of the present disclosure offer the possibility to determine a (global) network load in the multipath communication system in real time, which in turn can help to decide which scheduling scheme can guarantee the best quality of service in real time (e.g. at a specific time of data packet transmission) compared to the other existing scheduling schemes.

[0012] Some terms are used in this disclosure in the following ways: A "transmitter" (e.g., a transmitting unit) can be any data source in this disclosure. For example, in an uplink direction (i.e., the data packet transmission direction from the perspective of an end device toward a telecommunications network), a transmitter can be an audio, video, radio, TV, or sensor data transmitter (e.g., a mobile phone, computer, laptop, music player, or tablet) that can transmit data packets via appropriate transmission paths. In WLAN terminology, the transmitter can be referred to as a station or a transmitting node (e.g., a mobile phone, computer, or other device mentioned above equipped with a WLAN adapter or other network device). In some cases, the "transmitter" in WLAN terminology can be understood as a WLAN client of the station or transmitting node (e.g., an application).In the context of a mobile network, the term "transmitter" refers to a user equipment (UE) equipped with one or more mobile broadband adapters (e.g., a multi-SIM device) or other equipment. In some cases, the transmitter may operate in live streaming mode. Data packets (e.g., digital or analog signals corresponding to these data packets) may be output to a unit assigned to the selected transmission path on the transmitter side. Specifically, a data packet may be output to an internal or external network device of the transmitter assigned to the selected transmission path, in order to transmit the data packet to the receiver using that network device.

[0013] A "receiver" is a device located away from a transmitter that can receive data packets sent by a transmitter (e.g., directly via appropriate transmission paths). In the context of WLAN technology, in the present disclosure, a "receiver" refers to a wireless access point (ZP) in an uplink direction. In some cases, the term "receiver" may also include a base station of a mobile network in an uplink direction when data packets are transmitted via the mobile network.

[0014] In the case of the downlink direction (i.e., the data packet transmission direction coming from the perspective of an end device from the direction of a telecommunications network), the "receiver" in this disclosure will correspond to the "sender" of the uplink direction as defined above. Furthermore, the "sender" of the downlink direction will correspond to the "receiver" of the uplink direction as defined above.

[0015] The term "transmission path" refers to a communication path (or communication route) between a sender and a receiver (as defined above), enabling them to communicate with each other. The sender can be connected to the receiver via multiple (different) transmission paths, so a multi-path communication system comprises several transmission paths along with the sender and receiver. Two transmission paths can be spatially distinct, for example, if they run between a sender and receivers (e.g., two or more access points) located at different sites. In some cases, two transmission paths can differ in that they use different frequency channels or frequency bands for data packet transmission (e.g., 1 / 2 GHz).A transmitter sends data packets to a receiver over two or more frequency channels or frequency bands, resulting in the formation of two or more corresponding transmission paths. In some examples, two transmission paths differ due to the different radio access technologies used. In other examples, two different transmission paths can exhibit any combination of the differences mentioned above. Two different transmission paths can use identical or different communication protocols (e.g., WLAN protocols of the IEEE 802.11 family, such as 802.11ah or later) for data packet transmission. The radio connection can be a WLAN connection, a WPAN connection (e.g., using a transmission protocol of the IEEE 802.15.4 standard), or a cellular connection, such as UMTS, GPRS, 4G, LTE, or 5G.

[0016] The term "transmission path parameter" refers to any parameter relevant to the transmission of a data packet along a transmission path, or any parameter of the transmission path that influences the transmission of the data packet. According to the claim, the signal-to-noise ratio (SNR) of data packets on the transmission path is such a transmission path parameter. The queue length of the transmission path on the sender side is also a transmission path parameter within the meaning of the present invention. The queue length can be defined, for example, as the number of data packets in the queue (e.g., on the sender side) that are to be transmitted via the respective transmission path. A transmission latency (or simply latency) can also be a transmission path parameter.Within the scope of the present invention, transmission path latency can be understood as the time required to transmit the data packet along the transmission path from the sender to the receiver. Furthermore, jitter can be chosen as a transmission path parameter, which denotes a fluctuation in latency during transmission over a transmission path. Transmission path parameters can include, for example, a transmission data rate (defined, for example, as the number of units of information transmitted per unit of time), a transmission capacity, or a transmission bandwidth. Other, non-exhaustive, examples of transmission path parameters include a transmission path loss probability and a transmission path packet loss rate, the latter representing the number of data packets lost during transmission relative to the number of data packets sent.

[0017] "A correlation" between one or more transmission path parameters of one transmission path and one or more transmission path parameters of another transmission path within the scope of the present invention describes the extent to which one or more transmission path parameters of one transmission path (e.g., latency, signal-to-noise ratio, packet loss rate, or similar) change (e.g., over time) when data packets are transmitted via another transmission path characterized by a corresponding transmission path parameter. In other words, a correlation between different transmission paths can describe how the transmission path parameters of the different transmission paths are related over time (more on this below). In this sense, one can also speak of "the correlation" between two or more transmission paths.

[0018] Accordingly, the term "quality of service" encompasses the quality of a communication link (also referred to as "Quality of Service" or "QoS") over such a transmission path, which is described by a set of transmission path parameters defined above. Therefore, these QoS parameters can include, for example, data rate, transmission capacity, latency, transmission bandwidth, transmission reliability (e.g., maximum frame loss, maximum number of bit flips, maximum probability of CRC errors, need for redundant transmission), jitter, or a combination thereof. Furthermore, in this context, quality of service can also include other characteristics such as IT security and / or functional safety or reliability. IT security can refer to the protection of data (e.g., messages), among other things.Regarding integrity, this refers to the requirement that a message must not be altered in transit (e.g., during data transmission within a communication system and / or between the communication system and other systems). On the other hand, functional safety or reliability in the present technologies refers to operational safety (which can be compromised, for example, by excessive latency), i.e., the protection of people and the environment.

[0019] Accordingly, the term "communication network" encompasses the transmission paths defined above between one or more senders and corresponding one or more receivers. In this sense, these transmission paths and the resulting communication network can transmit data packets via one or more WLAN networks and / or one or more mobile networks. Thus, the communication network of the present invention can experience a (global) network load in connection with the transmission of data packets over multiple transmission paths, which can be characterized by a network parameter, e.g., a network load parameter (more on this below).

[0020] The term "multipath communication system" refers to any infrastructure for transmitting data packets over multiple transmission paths within that multipath communication system. Furthermore, the multipath communication system includes the communication network defined above. The multipath communication system of this disclosure may comprise subsystems, such as, as mentioned above, a sender (e.g., a mobile phone, a computer, or a laptop) and one or more receivers (e.g., one or more access points), which are connected by corresponding transmission paths of the multipath communication system and can communicate with each other. The communication system may also comprise two or more senders, which are connected by transmission paths of the multipath communication system to one or more respective receivers.Furthermore, receivers and / or the sender can transmit information to other devices or receive information or requests from other devices that are not part of the multipath communication system.

[0021] A scheduling scheme for data packet transmission in a multi-path communication system of the present invention determines which data packets (or their fragments) are transmitted via which transmission path (e.g., from a sender to a receiver, as discussed above). Furthermore, the scheduling scheme can define the sequence and / or data points to which data packets are transmitted via the communication system's transmission paths (e.g., from a sender to multiple receivers) (more on this below).

[0022] A "data packet" can, for example, contain video data, voice data, measurement data, or messages. The data packet can also contain coded data, such as data that has been encoded using an encoding technique. The data packet can be transmitted along the transmission path from the sender to the receiver. In the present disclosure, a plurality of data packets can include test data packets (i.e., data packets whose content is unimportant to the user) that can be used to adaptively select a scheduling scheme for data packet transmission in a multi-path communication system. Brief description of the characters

[0023] Fig. 1a is a flowchart that represents an example of an adaptive selection of a scheduling scheme for data packet transmission in a multipath communication system according to the first aspect. Fig. 1bis a flowchart that shows further possible process steps according to the first aspect. Fig. 2 Figure 1 schematically shows an exemplary structure of a multi-path communication system 1, comprising a transmitter 3, three wireless access points 4 (ZP1-ZP3) and corresponding transmission paths 2, via which the transmitter and the access points can exchange data with each other. Figs. 3a to 3c Figure 10 schematically illustrates three exemplary scheduling schemes for data packet transmission in a multipath communication system on the sender side. LB: Load Balancing; PD: Packet Duplication; PS: Packet Splitting. Fig. 4 schematically shows a flowchart 11 and further aspects of a procedure for adaptively selecting a scheduling scheme LB; PD; PS for data packet transmission in a multipath communication system 1. Fig. 5schematically shows an exemplary architecture for an implementation 12 of the scheduling schemes 10 with an adaptive multipath scheduler 5 and two transmission paths of a multipath communication system using an ns-3 simulator (see, for example, the following link). https: / / www.nsnam.org / ), which is applicable to the IEEE 802.11 standard. MAC: Media Access Control. PHY: Physical Layer. SNR: Signal-to-Noise Ratio. WL: Queue Length. CW: Contention Window. MCS: Modulation and Coding Scheme. Detailed description

[0024] First, using Fig. 1a and 1b Techniques for adaptively selecting a scheduling scheme for data packet transmission in a multipath communication system are described. An exemplary structure of a multipath communication system is then presented based on... Fig. 2 discussed. Next, the following will be discussed: Figs. 3a to 3c Three exemplary scheduling schemes for data packet transmission in a multipath communication system are shown. Then... Fig. 4 a possible design and further aspects of a procedure of the present disclosure will be illustrated. Finally, based on Fig. 5 An architecture for implementing scheduling schemes with an adaptive multipath scheduler is described.

[0025] As in the Fig. 1a and 1bAs outlined, a first general aspect concerns a method for adaptively selecting a scheduling scheme LB; PD; PS for data packet transmission in a multipath communication system 1. In the techniques presented here, adaptive selection of the scheduling scheme, as explained in more detail below, means that the scheduling scheme can be selected during data packet transmission (i.e., in real-time operation) such that this selected scheduling scheme (e.g., at a specific time) can ensure a better quality of service (e.g., lower latency or a lower probability of data packet loss) than the other available scheduling schemes. The specific examples of uplink data packet transmission discussed below serve to illustrate the application of the techniques presented here. However, the lesson contained therein is not limited to the uplink direction and can, for example, be applied to other directions.for data packet transmission in the downlink direction. The method steps of the corresponding independent claim are shown in the boxes drawn by solid lines in . Fig. 1a and 1b The process steps of some dependent claims are shown in the boxes represented by dashed lines.

[0026] The first step of the method comprises determining at least two transmission path parameters for each transmission path (e.g., a signal-to-noise ratio and a queue length on the sender side) of at least two transmission paths from the two or more transmission paths for a plurality of data packets. According to the claim, the at least two transmission path parameters comprise a respective queue length and a signal-to-noise ratio of data packets from the plurality of data packets that are transmitted via the corresponding transmission path of at least two transmission paths. In some cases, the at least two transmission path parameters for each transmission path (e.g., for the three in Fig. 2The majority of data packets are determined from the two or more transmission paths shown. In the present disclosure, the majority of data packets are transmitted via the two or more transmission paths of the multipath communication system. In some cases, the multipath communication system may include one or more transmitters 3 (e.g., one or more mobile phones and / or one or more computers) and one or more receivers 4 (e.g., one or more access points). In the example of Fig. 2A multi-path communication system 1 is schematically represented, comprising a transmitter 3, three receivers 4 (access points ZP1 to ZP3), and three transmission paths 2 through which the transmitter and receivers can be connected. In this example, the majority of data packets can be transmitted from the transmitter 3 to the respective receiver "4; ZP1-ZP3" via a transmission path "2; Path1-Path3". As mentioned above, the transmitter can be connected to the receiver via multiple (different) transmission paths. For example, one transmission path on the transmitter side can transmit 3 data packets to the first access point ZP1 via a WLAN connection using one frequency channel, and another transmission path on the transmitter side can transmit 3 data packets to the first access point ZP1 via the WLAN connection using a different frequency channel.In some cases, one transmission path on the sender side can transmit data packets via a WLAN connection, while the other two transmission paths can each be assigned to a mobile network (e.g., provided by two different mobile network operators), so that data packets are transmitted via a corresponding mobile connection (e.g., to a corresponding base station) (in . Fig. 2 (not shown).

[0027] In some cases, the transmission path can have a sender endpoint encompassed by the sender and a receiver endpoint encompassed by the receiver. The sender endpoint can be a protocol start point of a communication protocol (hereinafter referred to as "protocol sender") for transmitting the data packet from the sender to the receiver (e.g., from a transmitting unit to a receiving unit). The sender endpoint can be part of the sender, e.g., a network device of the sender. In the present disclosure, the receiver endpoint can be a protocol endpoint of a communication protocol (hereinafter referred to as "protocol receiver") for transmitting the data packet from the sender to the receiver. Furthermore, the receiver endpoint can be part of the receiver. The transmission path can preferably extend from the sender, in particular via or through the network device of the sender associated with the transmission path, to the protocol endpoint of the receiver. The sender 3 can, for example,The system can be configured to send a data packet or a plurality of data packets to receiver 4 via a respective interface. The receiver can, in turn, be configured to receive the data packet (e.g., via one or more respective interfaces). For the purposes of this disclosure, each pair of sending and receiving interfaces can be part of a respective transmission path.

[0028] In the present disclosure, the at least two transmission path characteristics of a transmission path of at least two transmission paths (e.g., a first transmission path, a second transmission path, or another transmission path different from the first and second transmission paths) can include a respective queue length and a signal-to-noise ratio (SRV) of data packets from the plurality of data packets transmitted over the transmission path of at least two transmission paths. In some cases, a packet loss rate or a contention window (CW) can be one of the transmission path characteristics, which is an indicator of packet loss. As mentioned above, latency, jitter, data rate, packet loss rate, transmission reliability, or a combination thereof can be the other transmission path characteristics of the transmission path. Returning to the example of Fig. 2 :The queue lengths of data packets on the sender side 3 and the signal-to-noise ratios of data packets transmitted from sender 3 to the three access points ZP1 to ZP3 can be determined for all three transmission paths "2; Path1-Path3". For example, the queue lengths can be determined on the sender side, while the signal-to-noise ratios can be determined on the receiver side, e.g., by a corresponding component of the sender (or receiver) or an external component connected to the sender (or receiver). In some cases, the receiver and the sender can exchange information about the transmission path characteristics (e.g., via a corresponding transmission path). Alternatively or additionally, the information about the transmission path characteristics can be transmitted to an adaptive multipath scheduler 5 (e.g., from the sender and / or receiver) and analyzed there (see...). Fig. 5 and further discussions).

[0029] Next, the present techniques comprise determining a correlation between at least one transmission path characteristic of at least two transmission path characteristics of a transmission path and a corresponding transmission path characteristic of at least two transmission path characteristics of one or more other transmission paths from the two or more transmission paths for the majority of data packets. As mentioned above, the correlation between the transmission path characteristics of different transmission paths can characterize the extent to which transmission path characteristics (according to claim 1, a queue length and an SNR, or (not specified in claim 1) one of the other transmission path characteristics mentioned above) of an individual transmission path depend on the transmission path characteristics of one or more other transmission paths. For example,One or more transmission path characteristics of a first transmission path change when the data packets are transmitted over a different transmission path (e.g., when the same sender sends the data to two different receivers via the two transmission paths). In one example, the correlation can be described using the Pearson correlation coefficient. In some cases, the transmission path characteristics of the transmission paths can change over time, so the correlation between the transmission path characteristics of the involved transmission paths can be a function of time.

[0030] The next step of the procedure involves the adaptive selection of a scheduling scheme for data packet transmission over two or more transmission paths, based on the specified transmission path characteristics and correlation. Three exemplary scheduling schemes for data packet transmission of the multipath communication system on the sender side are shown in Figs. 3a to 3c The diagram shows two transmission paths schematically. In the load balancing (LB) scheme, each packet is transmitted individually via only one transmission path (data packets 1 and 3 of the Fig. 3a are transmitted via "path 1", while data packets 2 and 4 are transmitted via "path 2". In the example of the packet duplication (PD) scheme, the data packet is replicated and transmitted via one transmission path, while its duplicate is transmitted via one or more other transmission paths (data packets 1 and 3 in Fig. 3bare transmitted via "path 1" and their duplicates via "path 2"). When the packet fragmentation scheme is applied to data packet transmission (or in other words, to the distribution of data packets across transmission paths), the data packet is fragmented and its fragments are transmitted via one or more other transmission paths (fragments 1.1 and 2.1 of data packets 1 and 2 in Fig. 3c are transmitted via "path 1", while fragments 1.2 and 2.2 of the same data packets 1 and 2 are transmitted via "path 2").

[0031] The method of the present disclosure comprises determining at least one network characteristic of a communication network of the multipath communication system. According to claim 1, the at least one network characteristic of the communication network of the multipath communication system (e.g., of the one described in claim 1) is... Fig. 2In the multipath communication system shown, a network load parameter describes a network load that can vary over time (more on this below). In some cases, information about the network load parameter can be transmitted from one or more receivers to the respective transmitters (e.g., to all transmitters) and / or to the multipath scheduler. For example, in the case of a WLAN connection (e.g., using the IEEE 802.11 standard), this information can be determined by an access point and transmitted to the respective transmitters (e.g., in an additional field of the beacon message) via beacon messages (also referred to as "beacon frames" in the IEEE 802.11 standard). Furthermore, the "adaptive selection" step can be performed based on at least one network characteristic (e.g., the network load parameter) of the communication network.

[0032] In the techniques presented, adaptive selection of the scheduling scheme for data packet transmission (i.e., in real-time operation as mentioned above) across the two or more transmission paths can occur according to a schedule. In one example, the schedule might stipulate that a predetermined number of data packets from the majority of data packets should be transmitted before adaptive selection is performed (i.e., if this condition is not met, the previously used scheduling scheme for data packet transmission continues to be used). In other examples, the schedule might stipulate that data packets from the majority of data packets should be transmitted within a predetermined time interval before adaptive selection is performed.

[0033] Alternatively or additionally, the adaptive selection of the scheduling scheme for data packet transmission across the two or more transmission paths can occur automatically following a predetermined trigger event. For example, the predetermined trigger event might involve one or more transmission path characteristics of one or more transmission paths meeting a transmission path characteristic criterion. For instance, the transmission path characteristic criterion might involve one or more transmission path characteristics (e.g., jitter, queue length, latency, packet loss rate, or a combination thereof) of one or more transmission paths exceeding a predetermined threshold. In other examples, the predetermined trigger event might involve one or more transmission path characteristics (e.g.,A predetermined trigger event might occur when a signal-to-noise ratio, data rate, or a combination thereof of one or more transmission paths falls below a predetermined threshold. In other examples, the predetermined trigger event might involve a change in one or more transmission path characteristics that meets a change criterion. For example, the change criterion might involve a change in one or more transmission path characteristics (e.g., a change in signal-to-noise ratio, queue length, latency, data rate, or a combination thereof) of one or more transmission paths exceeding a predetermined change threshold of the transmission path characteristic.

[0034] In other words, the predetermined trigger event can be defined based on a requirement for the quality of service described above (including quality of service related to IT security and / or functional safety or reliability) that one or more transmission paths of the multipath communication system should fulfill. In some cases, an application on the sending side that can (or wants to) communicate via the corresponding transmission paths can report a connection request to the adaptive multipath scheduler and / or to a corresponding component of the sending system (possibly specifying the desired or necessary QoS parameters). In some cases, the predetermined trigger event can contain information about a predetermined value of the network parameter (e.g., the network load parameter), above (or below) which adaptive selection should be performed.

[0035] In the present techniques, a first predetermined network characteristic criterion and a second predetermined network characteristic criterion can be assigned to the at least one network characteristic of the communication network. For example, the first predetermined network characteristic criterion 21 can include that the at least one network characteristic of the communication network falls below a first predetermined threshold. Furthermore, the second predetermined network characteristic criterion 22 can include that the at least one network characteristic of the communication network falls below a second predetermined threshold, where the second predetermined threshold is greater than the first predetermined threshold. As described above, a network load parameter, for example, can be selected as such a network characteristic.In some examples, the network load parameter can represent the network load in the communication network relative to the available network capacity. For example, the network load parameter could be the ratio between the network load and the available network capacity. The network load can be calculated at a specific point in time when data packets are transmitted by one or more senders using appropriate scheduling schemes, such as before adaptive selection of the first aspect is performed (a sender's scheduling scheme may differ from or be the same as another sender's). In other examples, a network load can be calculated as the average of network loads over a predetermined time interval.Furthermore, network capacity can represent the amount of data traffic over the transmission paths that the communication network can handle at a given time (e.g., data packets can be transmitted from one or more senders with a predetermined quality of service).

[0036] In some examples, the network load can be proportional to the number of senders. n in the multipath communication system, the network load can be proportional to the number of active senders (in other words, the senders involved in data packet transmission). Additionally, network load can be proportional to the average data rate at which senders (e.g., their applications) transmit data packets over the respective transmission paths. In some cases, network load can be proportional to an average number of retransmissions. ton the transmission paths of the multipath communication system, where the retransmission on a transmission path represents the number of retransmissions required when the data packet is sent on that transmission path. In some cases, the network capacity r can be proportional to an average of the sum of the transmission capacities of the transmission paths, where a transmission capacity of a transmission path (including, for example, the receiver and / or sender) represents a transmission rate that the transmission path can guarantee for data packet transmission. In one or more of the above cases, the averaging can be performed, for example, over several (e.g., all) senders of the multipath communication system involved in data packet transmission and / or over a corresponding time interval of data packet transmission. In an example, the network load parameter Ldefine as follows: L = α ⋅ s a ⋅ n b ⋅ β + t c r d where α, β, a, b, c and d The factors are predetermined. It is conceivable that these factors could take the following exemplary list of values: α = 1.1, β = 1, a = 1, b = 2, c = 1 and d = 1.

[0037] In the techniques presented, the "adaptive selection" step can include the adaptive selection of a first scheduling scheme "LB; 30a" for data packet transmission if at least one network characteristic does not satisfy the first predetermined network characteristic criterion 21 and does not satisfy the second predetermined network characteristic criterion 22. For example, the "adaptive selection" step can include the adaptive selection of the first scheduling scheme "LB; 30a" for data packet transmission if at least one network characteristic exceeds both the first predetermined threshold and the second predetermined threshold 21; 22 (see also the discussions above). As discussed above, the network load parameter, for example, can be selected as the network characteristic. In the example of Fig. 4The first predetermined network characteristic criterion 21 in the diamond of flowchart 11 can be formulated as an inequality with respect to the network load parameter that is smaller than the first predetermined threshold, while the first predetermined network characteristic criterion 22 in the diamond of flowchart 11 can be formulated as an inequality with respect to the network load parameter that is smaller than the second predetermined threshold. In the present disclosure, the first scheduling scheme "LB; 30a" can be the load balancing scheme (LB) introduced above if each packet is transmitted individually over only one transmission path (see Figure 30a). Fig. 3a ).In some cases, the load balancing scheme can be advantageous when the network load parameter exceeds the two predetermined thresholds, as this scheme can reduce traffic per transmission path by distributing data packets across different paths. Furthermore, the load balancing scheme can increase inter-packet times when necessary (e.g., doubling them, as in the round-robin method, for example, with cyclic packet transmission across the two paths) to reduce queue length on the sender side.

[0038] The techniques of the present disclosure may further include selecting a second scheduling scheme "PD; 31a" for data packet transmission if the at least one network characteristic does not satisfy the first predetermined network characteristic criterion 21 and satisfies the second predetermined network characteristic criterion 22 (e.g., if the at least one network characteristic exceeds the first predetermined threshold and is below the second predetermined threshold), and if the specified signal-to-noise ratio (SRV) of data packets of the transmission path (e.g., a first transmission path of a sender, a second transmission path of the sender, or another transmission path that differs from the first and second transmission paths of the sender, or a combination thereof) satisfies a first predetermined criterion 23 (this criterion is set out in Fig. 4(schematically represented in the diamond with reference number 23). In one example, the specified SRV of the transmission path can be an SRV average of the transmission path, calculated by averaging the signal-to-noise ratios of a given number of data packets (e.g., 10 or more, 50 or more, 100 or more) transmitted over that transmission path. In other examples, the specified SRV of the transmission path can be an SRV average of the transmission path, calculated by averaging the signal-to-noise ratios transmitted over that transmission path within a given time interval (e.g., 1 ms or more, 10 ms or more, 100 ms or more). In some cases, the signal-to-noise ratios of one or more (e.g., all) transmission paths extending from a sender to the respective receivers can be determined. In this way, for example,The signal-to-noise ratios for several (e.g., all) transmitters are determined.

[0039] In the techniques presented here, the first predetermined criterion can include the requirement that the signal-to-noise ratio (SRV) of data packets in the transmission path falls below an SRV threshold (see the following explanations for calculating the SRV threshold). Returning to the example of Fig. 2 : The signal-to-noise ratios, as defined above, can be determined for all three transmission paths "2; Path 1-Path 3". In this embodiment, the first predetermined criterion can state that the first predetermined criterion is met if the SRV of any one of these three transmission paths falls below the SRV threshold. In other examples, two or all of the three determined signal-to-noise ratios must fall below the SRV threshold to meet the first predetermined criterion.

[0040] Furthermore, the procedure of the first aspect can include calculating an SRV network characteristic, which characterizes the signal-to-noise ratios of data packets from the majority of data packets transmitted over the communication network of the multipath communication system. In some cases, the SRV network characteristic can be an SRV mean of the signal-to-noise ratios of the data packets from the majority of data packets transmitted over the communication network of the multipath communication system (or, in other words, over the transmission paths that comprise the communication network). This SRV mean, which represents a global SRV of the communication system, can be calculated, for example, as an average of the specific (described above) SRV mean values ​​of the multiple (e.g., all) transmission paths of the multipath communication system.

[0041] In the next step, the procedure from the first aspect may include calculating the SRV threshold based on the calculated SRV characteristic. In some cases, the SRV threshold can be calculated as the difference between the SRV network characteristic and a predefined SRV shift value. The predefined SRV shift value could, for example, be selected from a range of 0 dB to 5 dB. It is conceivable that the SRV shift value could take the following non-exhaustive list of values: 0.5 dB, 1.0 dB, 1.5 dB, 2.0 dB, 2.5 dB, 3.0 dB, 3.5 dB, 4.0 dB, or 4.5 dB.

[0042] In the present disclosure, the second scheduling scheme "PD; 31a" can be the packet duplication (PD) scheme introduced above, in which the data packet is replicated and transmitted over one transmission path, while its duplicate is transmitted over one or more other transmission paths (e.g., from the same sender). With this scheme, the receiver must wait for the first arriving duplicate of the data packet (or for the data packet itself if it arrives earlier), which can reduce latency. Furthermore, the receiver can fall back on the second transmitted duplicate if a data packet is lost, which can increase the reliability of data packet transmission.

[0043] In the other case, if at least one network characteristic does not meet the first predetermined network characteristic criterion 21 and meets the second predetermined network characteristic criterion 22 (e.g., if the at least one network characteristic exceeds the first predetermined threshold and is below the second predetermined threshold), and if the specific signal-to-noise ratio of data packets on the transmission path does not meet the first predetermined criterion 23 (e.g., with respect to the SRV threshold described in detail above), the procedure of the first aspect may include selecting the first scheduling scheme "LB; 30b" for data packet transmission. Here, the first scheduling scheme "LB; 30b" may, for example, be the load balancing scheme (LB) introduced above.

[0044] The techniques of the present disclosure may further include calculating a collective queue length that characterizes the determined queue lengths of data packets on the at least two transmission paths when the at least one network characteristic (e.g., the network load parameter introduced above) satisfies the first predetermined network characteristic criterion 21 (e.g., when the at least one network characteristic is below the first predetermined threshold). In one example, the collective queue length may be a weighted sum of the determined queue lengths of data packets (e.g., with weight factors equal to one) transmitted over the at least two transmission paths. In some cases, the collective queue length may be a sum of the determined queue lengths of data packets transmitted over the transmission paths (e.g.,All transmission paths are transmitted, extending from a sender to the respective receivers. In some cases, the queue length for a transmission path (e.g., for each transmission path) can be determined as a queue length within a predefined time interval (e.g., a maximum measured queue length within this time interval can be taken as the queue length for the respective transmission path). In this way, for example, the queue length for multiple (e.g., all) senders can be determined.

[0045] In the present techniques, the step "adaptive selection of the scheduling scheme" can include selecting the second scheduling scheme "PD; 31b" for data packet transmission if the calculated collective queue length satisfies a second predetermined criterion 24 (this criterion is in Fig. 4(schematically represented in the diamond with reference number 24), and if the determined correlation fulfills a third predetermined criterion 25 (this criterion is in Fig. 4(shown schematically in the diamond with reference number 25). In this context, in some cases, the correlation between one or more transmission path characteristics of a first transmission path and one or more transmission path characteristics of a second transmission path of the same sender, or between a first transmission path and another transmission path of the same sender that differs from the first and second transmission paths, can be determined. In still other cases, the correlations between one or more transmission path characteristics of all transmission paths that transmit data packets from the same sender to the receivers can be determined. A maximum correlation can then be taken from the determined correlations to check whether it fulfills the third predetermined criterion 25.

[0046] In the techniques of the present disclosure, the second predetermined criterion 24 may include that the calculated collective queue length falls below a predetermined queue threshold. Furthermore, the third predetermined criterion 25 may include that the specified correlation falls below a predetermined correlation threshold. As mentioned above, the correlation can be described using Pearson's correlation coefficient. The correlation threshold may, for example, be selected from an interval of -1.0 to 1.0. It is conceivable that the correlation takes the following non-exhaustive list of values: -0.9, -0.5, -0.2, -0.1, 0.0, 0.1, 0.2, 0.5, or 0.9. As discussed above, in some examples, the transmission paths of the same transmitter may be used to adaptively select the scheduling scheme for that transmitter. Furthermore, for example,The packet duplication scheme (PD) introduced above is selected as the second scheduling scheme "PD; 31b".

[0047] In the other case, if the calculated collective queue length satisfies the second predetermined criterion 24, and the determined correlation does not satisfy the third predetermined criterion 25 (e.g., if the determined correlation exceeds the predetermined correlation threshold), the procedure of the first aspect may include selecting a third scheduling scheme "PS; 32" for data packet transmission. In the present disclosure, the third scheduling scheme "PS; 32" may be the packet fragmentation scheme (PS) for data packet transmission introduced above, in which the data packet is fragmented and its fragments are transmitted via one or more other transmission paths (see Figure 350). Fig. 3cand the discussions above). In some cases, the PS scheme can be advantageous because it can reduce the size of the payload per transmission path, which can lead to reduced latency. For example, the size of the data packets transmitted over the two transmission paths in the example of Fig. 3c be halved.

[0048] The step "adaptive selection of the scheduling scheme" of the present disclosure may ultimately include the selection of the first scheduling scheme "LB; 30c" (e.g., the load balancing scheme (LB) operated above) for data packet transmission if the at least one network characteristic satisfies the first predetermined network characteristic criterion 21 (e.g., if the at least one network characteristic is below the first predetermined threshold) and if the calculated collective queue length does not satisfy the second predetermined criterion 24.

[0049] A second general aspect of the present disclosure relates to a computer program designed to execute the method according to the first general aspect of the present disclosure. The present disclosure also relates to a computer-readable medium (e.g., a machine-readable storage medium such as an optical storage medium or solid-state storage, e.g., flash memory) and signals that store or encode the computer program of the present disclosure.

[0050] A third general aspect of the present disclosure relates to an adaptive multipath scheduler 5 of a multipath communication system 1, which is designed to execute the method according to the first general aspect of the present disclosure and / or to execute the computer program according to the second general aspect.

[0051] In some examples, the decision as to which scheduling scheme of the first aspect is applied to which sender can be made in a Fig. 5 The multipath scheduler 5 shown is used, which employs various transmission path parameters from the lower MAC layers (e.g., MAC middle layer, MAC sublayer, or similar), such as queue length, SRV, or conflict window (e.g., current values ​​of these parameters). In this example, the MAC upper layer is assigned to all transmission paths (e.g., those extending from a sender to the respective receivers), while the lower MAC layers are assigned to individual transmission paths (e.g., two transmission paths 2, as in Fig. 5 (shown). In other examples, a different layer can be assigned to all transmission paths. The decision made by the multipath scheduler can be seen in the example of Fig. 5The MAC upper layer is used, and the individual data packets (and / or their fragments) can be distributed to the individual lower MAC layers according to the scheduling scheme. Furthermore, in some cases, the multipath scheduler can transmit information about the appropriate modulation and encoding scheme applied to a specific transmission path to the MAC lower layer.

[0052] A fourth general aspect of the present disclosure relates to a multipath communication system comprising an adaptive multipath scheduler 5 according to the third aspect. Furthermore, the multipath communication system according to the fourth aspect comprises a transmitter 3 designed to transmit a plurality of data packets to the receiver. The transmitter of the fourth aspect can be connected to the receiver via two or more transmission paths 2. The multipath communication system of the fourth aspect is designed to perform the method according to the first general aspect of the present disclosure.

Claims

1. Method for adaptively selecting a scheduling scheme (LB; PD; PS) for a data packet transmission in a multi-path communication system (1), the method comprising the following steps: determining (100) at least two transmission path characteristic variables for each transmission path of at least two transmission paths (2) from the two or more transmission paths for a plurality of data packets, the plurality of data packets being transmitted via the two or more transmission paths of the multi-path communication system (1); determining (200) a correlation between at least one transmission path characteristic variable of at least two transmission path characteristic variables of a transmission path and a corresponding transmission path characteristic variable of at least two transmission path characteristic variables of one or more other transmission paths from the two or more transmission paths for the plurality of data packets; adaptively selecting (300) a scheduling scheme (LB; PD; PS) for a data packet transmission via the two or more transmission paths on the basis of at least the determined transmission path characteristic variables and the correlation, the method further comprising determining (210) at least one network characteristic variable of a communication network of the multi-path communication system (1), the adaptive selection (300) furthermore taking place on the basis of the at least one network characteristic variable of the communication network, characterized in that the correlation describes the extent to which one or more transmission path characteristic variables of a transmission path change when data packets are transmitted via another transmission path, which is characterized by a corresponding transmission path characteristic variable, the at least two transmission path characteristic variables comprising a respective queue length and a signal-to-noise ratio (SNR) of data packets from the plurality of data packets that are transmitted via the corresponding transmission path of at least two transmission paths (2), and the at least one network characteristic variable of the communication network being a network load parameter that describes a network load.

2. Method according to Claim 1, the adaptive selection (300) of the scheduling scheme (LB; PD; PS) for the data packet transmission via the two or more transmission paths being performed according to a schedule (20) or automatically after a predetermined trigger event, a facet of the schedule being that a predetermined number of data packets from the plurality of data packets should be transmitted before the adaptive selection (300) is carried out, a facet of the predetermined trigger event being that one or more transmission path characteristic variables of one or more transmission paths meet a transmission path characteristic variable criterion.

3. Method according to Claim 1 or 2, the at least one network characteristic variable of the communication network being assigned a first predetermined network characteristic variable criterion and a second predetermined network characteristic variable criterion, a facet of the first predetermined network characteristic variable criterion (21) being that the at least one network characteristic variable of the communication network drops below a first predetermined threshold value, a facet of the second predetermined network characteristic variable criterion (22) being that the at least one network characteristic variable of the communication network drops below a second predetermined threshold value, the second predetermined threshold value being greater than the first predetermined threshold value.

4. Method according to Claim 3, the adaptive selection (300) of the scheduling scheme (LB; PD; PS) comprising adaptively selecting (310) a first scheduling scheme (LB; 30a) for the data packet transmission if the at least one network characteristic variable does not meet the first predetermined network characteristic variable criterion (21) and does not meet the second predetermined network characteristic variable criterion (22), the first scheduling scheme (LB; 30a) being a load distribution scheme (LB).

5. Method according to Claim 3, the at least two transmission path characteristic variables of a transmission path of at least two transmission paths (2) comprising a respective queue length and a signal-to-noise ratio (SNR) of data packets from the plurality of data packets that are transmitted via the transmission path of at least two transmission paths (2), the adaptive selection (300) of the scheduling scheme (LB; PD; PS) comprising the following steps if the at least one network characteristic variable does not meet the first predetermined network characteristic variable criterion (21) and meets the second predetermined network characteristic variable criterion (22): selecting (320) a second scheduling scheme (PD; 31a) for the data packet transmission if the determined signal-to-noise ratio of data packets of the transmission path meets a first predetermined criterion (23), and otherwise, if the determined signal-to-noise ratio of data packets of the transmission path does not meet the first predetermined criterion (23), selecting (330) a first scheduling scheme (LB; 30b) for the data packet transmission, the first scheduling scheme (LB; 30b) being a load distribution scheme (LB), the second scheduling scheme (PD; 31a) being a packet duplication scheme (PD).

6. Method according to Claim 5, a facet of the first predetermined criterion (23) being that the determined signal-to-noise ratio, SNR, of data packets of the transmission path drops below an SNR threshold value, and the method comprising the following steps: calculating an SNR network characteristic value that characterizes signal-to-noise ratios of data packets from the plurality of data packets transmitted via the communication network of the multi-path communication system (1); calculating the SNR threshold value on the basis of the calculated SNR characteristic value.

7. Method according to Claim 3, the at least two transmission path characteristic variables of a transmission path of at least two transmission paths (2) comprising a respective queue length and a signal-to-noise ratio (SNR) of data packets from the plurality of data packets that are transmitted via the transmission path of at least two transmission paths (2), the method comprising calculating a collective queue length that characterizes the determined queue lengths of data packets on the at least two transmission paths (2) if the at least one network characteristic variable meets the first predetermined network characteristic variable criterion (21).

8. Method according to Claim 7, the adaptive selection (300) of the scheduling scheme (LB; PD; PS) comprising the following steps if the calculated collective queue length meets a second predetermined criterion (24): selecting (340) a second scheduling scheme (PD; 31b) for the data packet transmission if the determined correlation meets a third predetermined criterion (25), and otherwise selecting (350) a third scheduling scheme (PS; 32) for the data packet transmission, a facet of the second predetermined criterion (24) being that the calculated collective queue length drops below a predetermined queue threshold value, a facet of the third predetermined criterion (25) being that the determined correlation drops below a predetermined correlation threshold value, the second scheduling scheme (PD; 31b) being a packet duplication scheme (PD), the third scheduling scheme (PS; 32) being a packet fragmentation scheme (PS).

9. Method according to Claim 8, the adaptive selection (300) of the scheduling scheme (LB; PD; PS) comprising selecting (360) a first scheduling scheme (LB; 30c) for the data packet transmission if the at least one network characteristic variable meets the first predetermined network characteristic variable criterion (21) and if the calculated collective queue length does not meet a second predetermined criterion (24), the first scheduling scheme (LB; 30c) being a load distribution scheme (LB), a facet of the second predetermined criterion (24) being that the calculated collective queue length drops below a predetermined queue threshold value.

10. Computer program designed to carry out the method according to one of preceding Claims 1 to 9.

11. Adaptive multi-path scheduler (5) of a multi-path communication system (1) designed to carry out the method according to one of preceding Claims 1 to 9 and / or to execute the computer program according to Claim 10.

12. Multi-path communication system (1) comprising: an adaptive multi-path scheduler (5) according to Claim 11; a transmitter (3) designed to transmit a plurality of data packets to the receiver; a receiver (4), the transmitter being connected to the receiver via two or more transmission paths (2), the multi-path communication system (1) being designed to carry out the method according to one of preceding Claims 1 to 9.