Techniques for Scheduling Multipath Data Traffic

Abstract

The present disclosure relates to a technique for scheduling data traffic for transmission via multiple paths, wherein at least one of the multiple paths is of a first type, specifically a cheap data path, and at least another path is of a second type, specifically an expensive data path. The present disclosure particularly relates to a multipath scheduler device (110) for scheduling data traffic (102) for transmission via at least one first type data path (111) and at least one second type data path (112), the multipath scheduler device (110) comprising: at least one first type data path (111); at least one second type data path (112); and a scheduler (113) configured to schedule data traffic (102) for transmission via the at least one first type data path (111), and to schedule data traffic (102) for delayed transmission via the at least one second type data path (112) when at least the second type data path 112 is requested.

Description

Technical Field

[0001] The present disclosure relates to techniques for scheduling data traffic for transmission over multiple paths, at least one of the multiple paths being of a first type, specifically a cheap data path, and at least another path being of a second type, specifically an expensive data path. The present disclosure also relates to a multipath scheduler and a method for scheduling multipath data traffic. The present disclosure particularly relates to detecting and balancing burst traffic in a multipath scenario to offload determined paths. Background Art

[0002] Several multipath protocols (e.g., MPTCP according to "TCP Extensions for Multipath Operation with Multiple Addresses" by A. Ford and C. Raiciu and M. Handley and O. Bonaventure, RFC No. 6824, January 2013", QUIC according to "QUIC: A UDP-Based Secure and Reliable Transport for Multipath QUIC for HTTP / 2 (QUIC: Secure and Reliable Transport over UDP for HTTP / 2), draft-tsvwg-quic-protocol-02, January 2016, Huawei's GRE Tunnel Binding Protocol, Leymann and C. Heidemann and M. Zhang and B. Sarikaya and M. Cullen, "Huawei's GRE Tunnel Binding Protocol", RFC No. 8157, May 2017, and many other protocols provide capacity aggregation of several paths. These paths can have different costs, such as cheap and expensive or expressed in other terms. For example, a cheap path can be a Wi-Fi path, while an expensive path can be an LTE (Long Term Evolution) path. Cost does not inevitably mean direct payment, it can also be defined in terms of latency, reliability and many other aspects.

[0003] A frequently used Internet service is non-real-time video streaming, which is usually of bursty nature. The video source attempts to fill the buffer at the well as quickly as possible and stops when the buffer is full. As long as the buffer is empty to a certain level, the source attempts to fill the well buffer again. In a multipath scenario, when additional capacity is provided, the video source will use all possible data paths to fill the well buffer as quickly as possible, and therefore the capacity of the cheap path can be combined with the capacity of the expensive path. This behavior may be undesirable for the customer, especially when the cheaper path is sufficient to meet the demand and does not significantly affect the QoE (quality of experience). Summary of the invention

[0004] It is an object of the present invention to provide a concept for efficiently scheduling data traffic, in particular bursty data traffic, in a multipath scenario where cheap and expensive paths are available.

[0005] The above-mentioned and other objects are achieved by the subject matter of the independent claims. Additional implementation forms are evident from the dependent claims, the description and the drawings.

[0006] The main idea presented in the present disclosure is to delay overflow traffic of cheaper paths (e.g., peak traffic that does not fit into cheaper paths) so that it can be filled into cheaper paths at a later time, for example, into the gaps between bursts of bursty data traffic, and thus avoid or at least limit the use of expensive data paths.

[0007] The methods and systems proposed below may be of various types. The various elements described may be implemented by hardware or software components, such as electronic components that can be manufactured by various technologies and include, for example, semiconductor chips, ASICs, microprocessors, digital signal processors, integrated electrical circuits, optoelectronic circuits and / or passive components.

[0008] The following proposed equipment, system and method can transmit information through a communication network. The term communication network refers to a technical infrastructure on which signal transmission is performed. The communication network basically includes a switching network in which the transmission and exchange of signals are performed between stationary devices and platforms in a mobile radio network or a fixed network, and an access network in which signal transmission is performed between a network access device and a communication terminal. The communication network may include both components of a mobile radio network and components of a fixed network. In a mobile network, an access network is also referred to as an air interface and includes, for example, a base station (B node, evolved B node, radio cellular cell) with a mobile antenna, which is used to establish communication to a communication terminal (e.g., a mobile phone or a mobile device with a mobile adapter) or a machine terminal as described above. In a fixed network, an access network includes, for example, a DSLAM (digital subscriber line access multiplexer) for connecting communication terminals of multiple participants based on a cable. Via a switching network, communication can be transferred to other networks, for example, other network operators, such as foreign networks.

[0009] In communications and computing systems, the Open Systems Interconnection model (OSI model) defines a conceptual model that characterizes and standardizes communication functions regardless of their underlying internal structure and technology. Its purpose is the interoperability of various communication systems under standard protocols. The model divides the communication system into abstract layers. The original version of the model defined seven layers: physical layer (layer 1), data link layer (layer 2), network layer (layer 3), transport layer (layer 4), session layer (layer 5), presentation layer (layer 6), and application layer (layer 7).

[0010] According to a first aspect, the present invention relates to a multipath scheduler device for scheduling data traffic for transmission via at least one first type data path and at least one second type data path, the multipath scheduler device comprising: at least one first type data path; at least one second type data path; and a scheduler configured to schedule data traffic for transmission via the at least one first type data path, and schedule the data traffic for delayed transmission via the at least one second type data path. In an exemplary implementation of the multipath scheduler device, a switch is arranged between the at least one first type data path and the at least one second type data path. In an exemplary implementation of the multipath scheduler device, the switch between the at least one first type data path and the at least one second type data path is based on a capacity threshold of the at least one first type data path and / or based on a traffic distribution function for distributing traffic between the at least one first type data path and the at least one second type data path. This concept (or exemplary implementation) should not be understood as a limitation, but should be understood as a descriptive example in the entire disclosure. Any kind of distribution functionality may be combined with delayed transmission via the at least one second type data path.

[0011] This multipath scheduler device can efficiently schedule data traffic, especially bursty data traffic, in a multipath scenario where cheap and expensive paths are available. Overflow traffic in a cheaper path (e.g., peak traffic that does not fit into a cheaper path) can be delayed so that it can be filled into the cheaper path at a later time, such as filling into the gaps between bursts of bursty data traffic.

[0012] In an exemplary implementation of a multipath scheduler device, transmission of data traffic via the at least one second type data path is more expensive than transmission of data traffic via the at least one first type data path, specifically in terms of latency, reliability, capacity, complexity and / or cost.

[0013] This provides the advantage that data traffic can be flexibly transmitted via different data paths with different costs, for example via a cheap Wi-Fi data path and an expensive LTE data path. Cost does not inevitably imply direct payment, it can also be defined in terms of latency, reliability and many other aspects.

[0014] In an exemplary implementation form of the multipath scheduler apparatus, the scheduler is configured to schedule the data traffic for delayed transmission via the second type of data path if a bursty traffic pattern of the data traffic is detected.

[0015] This provides the following advantages: when transmission is delayed in a bursty traffic situation, the peak traffic that would use the second type of data path (i.e., expensive data path) may fall between two peaks of the bursty data traffic and thus may be transmitted via the first type of data path (i.e., cheap data path). That is, the transmission cost may be reduced.

[0016] In an exemplary implementation form of the multipath scheduler apparatus, the scheduler is configured to detect the bursty traffic pattern based on a peak behavior of data traffic scheduled for transmission via the at least one second type data path.

[0017] This provides the advantage that bursty data traffic can easily be detected by inspecting the second type data path.

[0018] In an exemplary implementation form of the multipath scheduler device, the scheduler is configured to delay access to the at least one second type data path based on a trigger.

[0019] This provides the following advantages: The trigger can guarantee the same scheduling conditions. The trigger can be predetermined or can be adjusted depending on the traffic situation. Therefore, the multipath scheduler device can be adapted to different kinds of traffic situations.

[0020] In an exemplary implementation of the multipath scheduler device, the trigger is based on a traffic parameter or a combination of traffic parameters.

[0021] This provides the advantage that traffic parameters can be selected depending on the respective traffic situation, thereby allowing the triggering to be adjusted in order to optimally distribute the data traffic between the cheap and expensive data paths.

[0022] In an exemplary implementation of the multipath scheduler device, triggering the use of T based on the last path LPU and the corresponding path access delay T Delay .

[0023] This provides the advantage that the delay for bursty data traffic can be appropriately selected to provide appropriate use of cheap and expensive data paths.

[0024] In an exemplary implementation of the multipath scheduler device, the scheduler is configured to use T LPU When the value of the at least one second type data path is higher than a threshold, access to the at least one second type data path is delayed.

[0025] This provides the advantage that bursty traffic can be reliably detected and scheduling can be adjusted optimally.

[0026] In an exemplary implementation form of the multipath scheduler device, the scheduler is configured to Delay Access to the at least one second type data path is denied within a period of time.

[0027] This provides the advantage that a minimum delay during which the second type of data path (ie the expensive data path) is not used can be adjusted.

[0028] In an exemplary implementation of the multipath scheduler device, the trigger is based on the amount of data, the amount of data in a predefined time period, or a combination thereof.

[0029] This provides the advantage that the trigger can be optimally selected based on the traffic conditions.

[0030] In an exemplary implementation form of the multipath scheduler device, the triggering is based on information about a gap in traffic transmission on the at least one first type data path.

[0031] This provides the advantage that the trigger can be optimally selected based on information or knowledge about the gap in traffic transmission on the first type of data path (eg the cheaper data path).

[0032] In an exemplary implementation of the multipath scheduler device, the triggering is based on a heuristic method.

[0033] This provides the advantage that the trigger can be optimally selected based on applying such heuristics.

[0034] In an exemplary implementation form of the multipath scheduler device, the scheduler is configured to delay access to the at least one second type data path based on an access restriction function.

[0035] This provides the advantage that the access restriction function provides a more relaxed access to the second type data path than the delay function. Thus, a smoother access to the second type data path can be achieved.

[0036] In an exemplary implementation form of the multipath scheduler device, the access restriction function comprises a slope function, in particular an exponential function, a ramp function or a step function.

[0037] This provides the following advantages: various designs for implementing the access restriction function can be realized. Depending on the traffic situation, an appropriate access restriction function can be selected, thereby providing a high degree of flexibility.

[0038] In an exemplary implementation form of the multipath scheduler device, the scheduler is configured to buffer data traffic overflowing a capacity threshold and schedule the buffered data traffic for transmission via the at least one first type data path.

[0039] This provides the advantage that overflow data traffic can be buffered in a buffer and can be transmitted at a later time when the first type of data path is empty.Thus, transmission costs can be reduced by avoiding the use of expensive data paths.

[0040] In an exemplary implementation form of the multipath scheduler device, the at least one first type data path includes a transmit buffer configured to buffer overflow data traffic, wherein a size of the transmit buffer is adjustable based on the overflow data traffic.

[0041] This offers the advantage that by using such a buffer, depending on the buffer size, access to the second type of data path can be avoided or at least limited and the costs of data transmission can thus be reduced.

[0042] According to a second aspect, the present invention relates to a method for scheduling multipath data traffic for transmission via at least one first type data path and at least one second type data path, the method comprising: scheduling data traffic for transmission via the at least one first type data path and scheduling the data traffic for delayed transmission via the at least one second type data path. In an exemplary implementation of the method, the method further comprises switching between the at least one first type data path and the at least one second type data path. In an exemplary implementation of the method, switching between the at least one first type data path and the at least one second type data path is based on a capacity threshold of the at least one first type data path and / or based on a traffic distribution function. This concept (or exemplary implementation) should not be understood as a limitation, but rather as a descriptive example in the entire disclosure. Any kind of distribution function can be combined with delayed transmission via the at least one second type data path.

[0043] This method can efficiently schedule data traffic, especially bursty data traffic, in multipath scenarios where cheap and expensive paths are available. Overflow traffic in the cheaper path (e.g., peak traffic that does not fit into the cheaper path) can be delayed so that it can be filled into the cheaper path at a later time, such as filling into the gaps between bursts of bursty data traffic.

[0044] According to a third aspect, the invention relates to a computer-readable non-transitory medium having stored thereon computer instructions which, when executed by a computer, cause the computer to perform a method according to the second aspect of the invention.

[0045] According to a fourth aspect of the present invention, the present invention relates to a computer program product comprising program code for performing the method according to the second aspect of the present invention when executed on a computer or a processor.

[0046] The embodiments of the present invention may be implemented via hardware or software.

[0047] The following acronyms apply in this disclosure:

[0048] GRE Generic Routing Encapsulation

[0049] LPU Last used

[0050] MPTCP Multipath TCP

[0051] OSI Open Systems Interconnection

[0052] QUIC Fast UDP Internet Connection

[0053] TCP Transmission Control Protocol

[0054] UDP User Datagram Protocol BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Additional embodiments of the present invention will be described with reference to the following drawings, in which:

[0056] Figure 1 An exemplary multi-path communication system 100 having two paths of different types (specifically, different costs) according to the present disclosure is shown;

[0057] Figure 2 shows a throughput graph 200 illustrating an exemplary traffic pattern in a multipath scenario according to the present disclosure, the exemplary traffic pattern requiring spillover from one path to another;

[0058] Figure 3 A throughput graph 300 illustrating an ideal transfer 300 that overflows traffic into a less expensive path in accordance with the present disclosure is shown;

[0059] Figure 4 shows a throughput graph 400 illustrating peak detection and path usage delay on expensive paths in accordance with the present disclosure;

[0060] Figure 5 shows a throughput graph 500 illustrating peak detection and stretched slope behavior of expensive paths in accordance with the present disclosure;

[0061] Figure 6a , 6b and 6c show throughput graphs 600a, 600b, 600c illustrating a burst buffer on a cheaper path according to the present disclosure, the burst buffer being used to accept overflow traffic to a certain extent; and

[0062] Figure 7 A schematic diagram illustrating a method 700 for scheduling multipath data traffic for transmission via at least one first type data path and at least one second type data path according to the present disclosure is shown. DETAILED DESCRIPTION

[0063] In the following detailed description, reference is made to the accompanying drawings, which form a part of the present disclosure and in which specific aspects of the present invention may be placed are shown by way of illustration. It will be appreciated that other aspects may be used and structural or logical changes may be made without departing from the scope of the present invention. Therefore, the following detailed description is not intended to be limiting, as the scope of the present invention is defined by the appended claims.

[0064] For example, it should be understood that disclosures related to the method may also be true for a corresponding device or system configured to perform the method, and vice versa. For example, if a specific method step is described, the corresponding device may include a unit that performs the described method step, even if such a unit is not explicitly described or illustrated in the drawings. In addition, it should be understood that the features of the various exemplary aspects described herein may be combined with each other unless otherwise specifically stated.

[0065] Figure 1 An exemplary multipath communication system 100 having two paths of different types (specifically, different costs) is shown. On the transmitting side, a generator 101 is used to generate data traffic 102 for a scheduler 113 of a multipath scheduler device 110. The scheduler 113 schedules the data traffic 102 for transmission via a first type of data path (1) 111 associated with a cheap pipe and via a second type of data path (2) 112 associated with an expensive data pipe. On the receiver side, both data paths 111, 112 are combined for reception at the receiver 103.

[0066] Figure 1 A general approach for multipath usage is depicted, where each of the above-mentioned protocols (i.e., MPTCP, Multipath QUIC, and Huawei's GRE Tunnel Binding Protocol) can be applied. Figure 1 A distinction is made between two paths 111, 112 with different costs (cheap, expensive). Cost does not inevitably mean a direct payment, it can also be defined by latency, reliability and many other aspects. A single OSI layer 3 or OSI layer 4 connection can already benefit from increased capacity. A frequently used service on the Internet is non-real-time video streaming. Common traffic patterns are therefore bursty and have gaps in between when the capacity on the path is not a bottleneck. The video source tries to fill the buffer at the well as quickly as possible and stops when the buffer is full. As long as the buffer is empty to a certain level, the source tries again to fill the well buffer as quickly as possible.

[0067] In order to solve or alleviate the above technical problems, the present disclosure proposes a solution based on delaying the use of expensive paths. In addition, three different exemplary implementations are proposed later.

[0068] The concept is not limited to Figure 1112 is shown for illustration purposes, it can handle any number of paths as long as there is information about the path costs. Whenever the expensive path 112 is requested to carry traffic, an attempt is made to detect bursty traffic patterns, for example by identifying peak behavior on the expensive path 112. If so, access to the expensive path 112 is delayed. If longer requests to the expensive path 112 are observed, access is granted because the peak is no longer set or the less expensive path 111 cannot meet the capacity demand. This design can be used with Figure 1 The multipath scheduler 110 is shown to be directly related, but may also be applied externally.

[0069] Hereinafter, the multipath scheduler device 110 is described in more detail.

[0070] The multipath scheduler device 110 may be used to schedule the data traffic 102 for transmission via at least one first type of data path (eg, Figure 1 111) and at least one second type data path (e.g., Figure 1 The multipath scheduler device 110 includes: at least one first type data path 111; at least one second type data path 112; and a scheduler 113. The scheduler 113 is configured to schedule the data traffic 102 for transmission via the at least one first type data path 111 and schedule the data traffic 102 for delayed transmission via the at least one second type data path 112. In an exemplary implementation, the scheduler 113 may be configured to schedule the data traffic 102 based on a capacity threshold 201 (e.g., Figure 2 6) and / or scheduling data traffic 102 for delayed transmission via the at least one second type data path 112 based on the traffic distribution function. The use of the capacity of the at least one first type data path 111 as a threshold for overflow to the at least one second type data path 112 has a descriptive nature throughout the present disclosure. Any kind of distribution function can be combined with delayed transmission via the at least one second type data path.

[0071] Transmission of data traffic 102 via the at least one second type data path 112 is more expensive than transmission of data traffic 102 via the at least one first type data path 111 , in particular in terms of latency, reliability, capacity, complexity and / or cost.

[0072] The scheduler 113 may be configured to schedule the data traffic 102 for delayed transmission via the second type data path 112 if a bursty traffic pattern of the data traffic 102 is detected. The scheduler 113 may be configured to detect the bursty traffic pattern based on peak behavior of the data traffic 102 scheduled for transmission via the at least one second type data path 112.

[0073] The scheduler 113 may be configured to delay access to at least one second type data path 112 based on a trigger. The trigger may be based on a traffic parameter or a combination of traffic parameters. The trigger may be based on the last path using T LPU 402 (for example, as shown below Figure 4 described) and the corresponding path access delay T Delay 403 (see Figure 4 ). The scheduler 113 may be configured to use T LPU 402 is higher than the threshold value, delaying access to at least one second type data path 112 (for example, as described below with reference to Figure 4 The scheduler 113 may be configured to schedule the access delay T Delay 403, access to at least one second type data path 112 is denied (for example, as described below with reference to Figure 4 The trigger may be based on the amount of data, the amount of data in a predefined time period, or a combination thereof.

[0074] The scheduler 113 may be configured to delay access to at least one second type data path 112 based on the access restriction function 503 (eg, as described below with reference to Figure 5 The access restriction function 503 may include a slope function 503, specifically an exponential function 503 (e.g., as described below with reference to Figure 5 described), ramp function or step function.

[0075] The scheduler 113 may be configured to buffer the data traffic 102b that overflows the capacity threshold 201 and schedule the buffered data traffic 102b for transmission via at least one first type data path 111 (eg, as described below with reference to Figure 6a , 6b and 6c). The at least one first type data path 111 may include a transmit buffer 601 configured to buffer overflow data traffic 102b (eg, as described below with reference to Figure 6a , 6b, 6c). The size of the transmit buffer 601 may be adjusted based on the overflow data traffic 102b. The use of the capacity of at least one first type data path 111 as a threshold for overflowing to at least one second type data path 112 is of a descriptive nature throughout this disclosure. Any kind of distribution function may be combined with delayed transmission via the at least one second type data path.

[0076] Figure 2 A throughput graph 200 is shown illustrating an exemplary traffic pattern in a multipath scenario according to the present disclosure that requires spillover from one path to another.

[0077] As mentioned above Figure 1 A frequently used service on the Internet is non-real-time video streaming. The common traffic pattern is therefore bursty with gaps in between when the capacity on the path is not a bottleneck. A similar traffic pattern is Figure 2 In the case of a multipath scenario where extra capacity is provided, bursty traffic is induced to use this extra capacity and eventually the capacity of the cheaper path 111 is combined with the capacity of the expensive path 112.

[0078] exist Figure 2 In the example of , the past time is on the right side of the time axis, and the current time is shown as the origin of the diagram. The traffic is bursty, where the first burst includes basic traffic 102a and peak traffic 102b, and the second burst includes basic traffic 102c and peak traffic 102d. The capacity threshold 201 represents the boundary between basic traffic and peak traffic. Each burst of basic traffic 102a, 102c is filling the cheap pipe 111, and each burst of peak traffic 102b, 102d is filling the expensive pipe 112.

[0079] Figure 3 A throughput diagram 300 is shown illustrating an ideal transfer 300 of overflowing traffic into a less expensive path according to the present disclosure. In the diagram, past time is on the right side of the time axis, while the current time is shown as the origin of the diagram. The traffic corresponds to Figure 2 The traffic shown is bursty, wherein the first burst includes basic traffic 102a and peak traffic 102b, and the second burst includes basic traffic 102c and peak traffic 102d. Peak traffic 102b, 102d is transferred from expensive pipe 112 to cheap pipe 111. The transferred data traffic is represented by reference numerals 102e and 102f. Capacity threshold 201 represents the boundary between basic traffic and peak traffic.

[0080] In this example, the capacity of the cheaper path 111 is sufficient and the overflow 102b, 102d is loaded into Figure 3Gaps between bursts are shown. In principle, the bursty traffic pattern is not limited to non-real-time video streaming services, this is just an example.

[0081] The trigger for delaying access to the expensive path 112 may be determined by different parameters or combinations of parameters. Some examples are presented below, but these examples do not include every possibility of setting and obtaining trigger points and how to delay access to an expensive path.

[0082] Figure 4 A throughput diagram 400 illustrating peak detection and path usage delay on expensive paths according to the present disclosure is shown. In this diagram, past time is also on the right side of the time axis, while the current time is shown as the origin of the diagram. The traffic corresponds to Figure 2 and 3 The traffic shown is bursty, wherein the first burst includes basic traffic 102a and peak traffic 102b, and the second burst includes basic traffic 102c and peak traffic 102d. Peak traffic 102d is transmitted with a delay time T Delay The data traffic is at least partially diverted 401 from the expensive pipe 112 to the cheap pipe 111 in the case of 403. The at least partially diverted data traffic is indicated by reference numeral 102g. The capacity threshold 201 indicates the boundary between basic traffic and peak traffic.

[0083] exist Figure 4 The implementation shown specifies two parameters, and the final path uses (T LPU )402 and the corresponding path access delay (T Delay ) 403. Based on the last usage of the expensive path 112 compared to the current path request, ensure the burst traffic mode. LPU If the threshold is higher than the defined threshold, access to the path is denied for a period of time T Delay This will give the cheaper path 111 the possibility to carry a larger portion of the overall request if capacity is available. LPU It is not mandatory and can be replaced by any other trigger mechanism. For example, it can use information about the amount of data or the amount of data in a defined time period or a combination. In addition, it can use information about traffic transmission gaps on the cheaper path 111, or in general it can use heuristic methods.

[0084] Figure 5 A throughput graph 500 illustrating peak detection and stretched slope behavior of expensive paths in accordance with the present disclosure is shown. Figure 5 The mechanism shown is similar to Figure 4 The mechanism shown, but does not define T Delay , but replace it with a slowly removed access restriction 503 (e.g., an exponential function or a ramp function or an on / off function).LPU 402 can also be used as a trigger point for this limitation. This scenario can also be implemented by applying a special congestion control algorithm with a reduced slope. In an exemplary implementation, parameters can be defined to limit the stretching slope function in time, similar to T Delay 403.

[0085] Figure 6a , 6b 600a, 600b, 600c illustrate throughput diagrams of a burst buffer on a cheaper path according to the present disclosure, which is used to accept overflow traffic to a certain extent. In these diagrams, past time is also on the right side of the time axis, and the current time is shown as the origin of the diagram. The traffic corresponds to Figures 2 to 5 The traffic shown is bursty, wherein the first burst shown includes basic traffic 102a and peak traffic 102b.

[0086] Reference Figure 6a , b, c, another concept is applied, which focuses on buffering the overflow traffic 102b and transmitting the overflow traffic via the cheaper path 111 before applying scheduling. Figure 6a A high traffic demand (burst) that may overflow into the expensive path 112 is shown. Figure 6b The idea shown is to squeeze / transfer 602 the overflow part 102b into the send buffer 601 of the cheap path 111, and increase the send buffer 601 to buffer the overflow part 102b when necessary. Another common term for this buffer 601 is a burst buffer. The buffer can be a network protocol inherent buffer, a device buffer, an interface buffer or any other buffer, so as long as the buffer is arranged before the bottleneck of the cheap path 111 (such as Figure 6c This results in a higher fill level of the transmit buffer 601, since it suppresses the scheduled traffic on the path 111. The size of the transmit buffer 601 (respectively the size extension for unloading the transmit buffer 601) is similar to Figure 4 and 5 T Delay However, transferring the overflow traffic 102b to the transmit buffer has a higher risk of end-to-end delay.

[0087] Figure 7 A schematic diagram illustrating a method 700 according to the present disclosure for scheduling multipath data traffic for transmission via at least one first type data path and at least one second type data path. The data traffic may correspond to Figure 1 6 shows data traffic 102, 102a, 102b, 102c, 102d, 102e, 102f, 102g.

[0088] The method 700 includes scheduling 701 data traffic 102 for transmission via at least one first type data path 111 (eg, as described above with reference to Figure 1 The method 700 also includes: scheduling 702 the data traffic 102 for transmission via at least one second type of data path 112 (eg, as described above with reference to FIG. 6 ) if at least one second type of data path 112 is requested. Figure 1 to 6) for delayed transmission.

[0089] Method 700 may be used as described above. Figure 1 The multipath scheduler device 110 described in FIG. 6 is implemented.

[0090] Method 700 may include additional steps, such as, for example, Figure 1 The calculation blocks described above are Figure 2 Steps to 6 describe the mechanism.

[0091] The method 700 may be implemented on a computer. Another aspect of the present invention relates to a computer-readable non-transitory medium having computer instructions stored thereon, which when executed by a computer cause the computer to perform the method 700 .

[0092] Another aspect of the present invention relates to a computer program product including program code for performing the method 700 or the above functions when the program code is executed on a computer or a processor. The method 700 may be implemented as a program code that may be stored on a non-transitory computer medium. The computer program product may implement the above Figure 1 The techniques described in 6.

[0093] Although a particular feature or aspect of the present disclosure may be disclosed only with respect to one of several implementations or embodiments, such a feature or aspect may be combined with one or more other features or aspects of other implementations or embodiments, which may be desirable and advantageous for any given or specific application. In addition, to the extent that the terms "include", "have", "with" or other variations thereof are used in the specific embodiments or claims, such terms are intended to be included in a manner similar to the term "comprise". Similarly, the terms "exemplary", "for example" and "eg" are merely meant as examples, not the best or optimal. The terms "coupled" and "connected", along with derivatives, may have been used. It should be understood that these terms may have been used to indicate that two elements cooperate or interact with each other, whether they are in direct physical or electrical contact, or they are not in direct contact with each other.

[0094] Although specific aspects have been illustrated and described herein, those skilled in the art will appreciate that various alternative and / or equivalent implementations may be substituted for the specific aspects shown and described without departing from the scope of the present disclosure. This application is intended to cover any adjustments or changes to the specific aspects discussed herein.

[0095] Although elements in the following claims are described in a specific order, these elements are not necessarily intended to be limited to being implemented in that specific order unless the claim description otherwise implies a specific order for implementing some or all of the elements.

[0096] According to the above teachings, many substitutions, modifications and variations will be apparent to those skilled in the art. Of course, it will be readily appreciated by those skilled in the art that, in addition to those applications described herein, the present invention has many applications. Although the present invention has been described with reference to one or more specific embodiments, it will be appreciated by those skilled in the art that many changes may be made to it without departing from the scope of the present invention. Therefore, it should be understood that, within the scope of the appended claims and their equivalents, the present invention may not be practiced as specifically described herein.

Claims

1. A multipath scheduler device (110) for scheduling data traffic (102) for transmission via at least one first type data path (111) and at least one second type data path (112), the multipath scheduler device (110) comprising: at least one first type data path (111); at least one second type data path (112); as well as a scheduler (113) configured to schedule a portion of the data traffic (102) for transmission via the at least one first type data path (111) and to schedule a portion of the data traffic (102) for delayed transmission via the at least one second type data path (112) when a bursty traffic pattern of the data traffic is detected, wherein the portion of the data traffic scheduled for delayed transmission via the at least one second type data path is at least a first portion of the data traffic that overflows a capacity threshold of the at least one first type data path, and wherein when the bursty traffic pattern is detected, the scheduler (113) is further configured to: delaying access to the second type of data path (112) based on a trigger or access restriction function, and buffering and scheduling at least a second portion of the data traffic that overflows the capacity threshold of the at least one first type data path for transmission via the at least one first type data path (111) at a later time, The access restriction function comprises a slope function (503), a ramp function or a step function.

2. The multipath scheduler device (110) according to claim 1, The scheduler (113) is configured to schedule a portion of the data traffic (102) for delayed transmission via the at least one second type data path (112) based on a traffic distribution function for distributing the data traffic between the at least one first type data path (111) and the at least one second type data path (112).

3. The multipath scheduler device (110) according to claim 1 or 2, The transmission of data traffic (102) via the at least one second type data path (112) is more expensive than the transmission of data traffic (102) via the at least one first type data path (111), in particular in terms of latency, reliability, capacity, complexity and / or cost.

4. The multipath scheduler device (110) according to claim 1, The scheduler (113) is configured to detect the bursty traffic pattern based on peak behavior of data traffic (102) scheduled for transmission via the at least one second type data path (112).

5. The multipath scheduler device (110) according to claim 4, The trigger is based on a traffic parameter or a combination of traffic parameters.

6. The multipath scheduler device (110) according to claim 4 or 5, The trigger is based on the last path using T LPU (402) and the corresponding path access delay T Delay (403), The last path uses T LPU (402) refers to the difference between the time point when the traffic throughput of the first burst drops from the peak value to the capacity threshold and the time point when the traffic throughput of the second burst rises from the basic traffic to the capacity threshold, The path access delay T Delay (403) refers to the time required for the throughput of the second burst to rise from the capacity threshold to the peak value.

7. The multipath scheduler device (110) according to claim 6, The scheduler (113) is configured to use T LPU (402) delaying access to the at least one second type data path (112) when the data rate is higher than a threshold.

8. The multipath scheduler device (110) according to claim 7, The scheduler (113) is configured to Delay (403) denies access to the at least one second type data path (112) within a time period.

9. The multipath scheduler device (110) according to claim 5, The trigger may be based on the amount of data, the amount of data in a predefined time period, or a combination thereof, and / or wherein the triggering is based on information about a gap in traffic transmission of the at least one first type data path (111), and / or The triggering is based on a heuristic approach.

10. The multipath scheduler device (110) according to claim 1, Wherein the slope function (503) comprises an exponential function.

11. The multipath scheduler device (110) according to claim 1, The at least one first type data path (111) comprises a transmit buffer (601) configured to buffer overflow data traffic (102b), wherein a size of the transmit buffer (601) is adjustable based on the overflow data traffic (102b).

12. A method (700) for scheduling multipath data traffic (102) for transmission via at least one first type data path (111) and at least one second type data path (112), the method (700) comprising: scheduling (701), by a scheduler (113), a portion of the data traffic (102) for transmission via the at least one first type data path (111), and scheduling a portion of the data traffic (102) for delayed transmission via the at least one second type data path (112) when a bursty traffic pattern of the data traffic is detected, wherein the portion of the data traffic scheduled for delayed transmission via the at least one second type data path is at least a first portion of the data traffic that overflows a capacity threshold of the at least one first type data path, And wherein when the burst traffic mode is detected, the method further comprises: delaying access to the second type of data path (112) based on a trigger or access restriction function, and buffering and scheduling at least a second portion of the data traffic that overflows the capacity threshold of the at least one first type data path for transmission via the at least one first type data path (111) at a later time, The access restriction function comprises a slope function (503), a ramp function or a step function.

Images