Data transmission system and method used in computer

Abstract

The invention relates to the technical field of computer networks, in particular to a data transmission system and method used in a computer. According to the invention, a burst traffic early warning mechanism based on the data cache queue state is established, and the burst traffic, the overload traffic and the stable traffic in the network can be more accurately judged in combination with the real-time monitoring of the queue length gradient and the traffic condition label. According to the invention, the dynamic marking module adjusts the marking position and the marking intensity according to the flow condition label, and reasonably controls the flow adjustment position and force of the data packet; through accurate mark adjustment, timely flow control in a network is facilitated, too early or too late pause of transmission of a data packet is avoided, the stability and throughput of data transmission are improved, and resource waste and overload phenomena are avoided; by optimizing the maximum length of the queue and the overflow processing and marking mechanism, the problem of queue oscillation in the traditional network is solved, and the stability of the queue is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of computer networks, and in particular to a data transmission system and method used in a computer. Background Art

[0002] With the rapid development of network communication technology, especially in data transmission and network resource management, more and more computer systems and applications are facing the challenge of high-speed data traffic processing. Especially in large-scale distributed systems and cloud computing environments, problems such as network congestion and resource waste are particularly prominent. Existing computer data transmission systems mainly rely on static queue management and flow control strategies. Although these methods can effectively handle data transmission under low load conditions, they still have many shortcomings when facing burst traffic and high load scenarios.

[0003] In traditional computer networks, the transmission of data streams usually relies on fixed queue management and flow control strategies. However, in practical applications, network loads are highly dynamic and uncertain. The occurrence of burst traffic, network congestion, and traffic fluctuations often leads to queue overflow, packet loss, and decreased bandwidth utilization, thus affecting the overall system performance. Although existing scheduling algorithms based on queue length and traffic models can partially alleviate these problems, they rely on static threshold settings and lack the ability to dynamically respond to real-time network status.

[0004] In view of the above problems, it is necessary to propose a data transmission system and method for use in a computer. Summary of the invention

[0005] The purpose of the present invention is to solve the problems existing in the background technology and to provide a data transmission system and method used in a computer.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] In a first aspect, the present invention provides a data transmission system for use in a computer, comprising a burst traffic monitoring module, a dynamic marking module, a transmission control module and a queue management module.

[0008] The burst traffic monitoring module establishes a burst traffic warning mechanism based on the data cache queue status.

[0009] The burst traffic warning mechanism dynamically monitors the changes in the data cache queue and outputs the corresponding traffic situation label. The specific process is as follows:

[0010] The dynamic monitoring is achieved by analyzing the queue length and the queue length gradient;

[0011] The queue length refers to the number of data packets or data segments waiting to be processed or transmitted in each data cache queue;

[0012] The queue length gradient refers to the change in the queue length of each data cache queue over time;

[0013] As a preferred embodiment of the present invention, the flow status is detected according to the queue length and the queue length gradient, and the flow status label corresponding to each data cache queue is output. The corresponding relationship is:

[0014] If the queue length of a data cache queue is less than the minimum threshold and the queue length gradient is less than the minimum threshold, the low traffic label of the data cache queue is output;

[0015] If the queue length of a data cache queue is greater than the maximum threshold and the queue length gradient is less than the minimum threshold, a high traffic label of the data cache queue is output;

[0016] If the queue length of a data cache queue is less than the maximum threshold and greater than the minimum threshold and the queue length gradient is greater than the maximum threshold, then the burst traffic label of the data cache queue is output;

[0017] If the queue length of a data cache queue is greater than the maximum threshold and the queue length gradient is greater than the maximum threshold, the overload traffic label of the data cache queue is output;

[0018] Otherwise, the stable traffic label of the output buffer queue.

[0019] Each traffic condition label corresponds to a preset label strength and label position. The label position points to a specific position in the data cache queue, and the specific position corresponds to a data packet in the data cache queue.

[0020] The queue length gradient calculation and flow condition state detection are performed on all data cache queues at preset time intervals to obtain flow condition labels of all data cache queues.

[0021] The dynamic marking module identifies the traffic condition label identified by the burst traffic monitoring module, and selects a corresponding marking adjustment strategy, including marking position adjustment and marking strength adjustment.

[0022] The specific process of adjusting the mark position is as follows:

[0023] Set the mark position of the low-traffic tag to always be at the front of the data cache queue.

[0024] Set the tag position of the high-traffic tag to always be at the back end of the data cache queue.

[0025] Set the burst traffic tag position to always be in the middle of the data cache queue.

[0026] Set the mark position of the overload traffic label to always be at the front of the data cache queue.

[0027] No additional adjustment is made to the marking position of the stable flow label.

[0028] The marking strength adjustment comprehensively considers the overall fluctuation degree of the historical queue length of the data cache queue, the single fluctuation degree and the time difference from the single fluctuation moment to the current time; comprehensively considers the overall fluctuation degree of the historical queue length gradient, the single fluctuation degree and the time difference from the single fluctuation moment to the current time.

[0029] Specifically, the overall fluctuation degree is the variance, and the single fluctuation degree is the difference between the queue length and the queue length gradient and their own average values.

[0030] As a preferred embodiment of the present invention, the labeling intensity is allowed to decay over time.

[0031] The transmission control module controls the data transmission rate and performs queue management. It adjusts the data transmission rate according to different network conditions to ensure maximum throughput and avoid network congestion and resource waste.

[0032] The data transmission rate control comprehensively considers the transmission rate adjustment base of each data cache queue, and the adjustment amount is inversely proportional to the sum of the marking strengths of each data cache queue and inversely proportional to the network transmission delay at the current moment.

[0033] The queue management described is related to the type of traffic condition label.

[0034] Specifically, for the data cache queue with high traffic labels, its transmission rate is increased and the transmission window size is adjusted;

[0035] For the data cache queue of low-traffic tags, the transmission rate is kept unchanged to avoid unnecessary waste of resources;

[0036] For the data buffer queue with burst traffic labels, adjust the congestion window.

[0037] The queue management module adjusts the maximum length of the data cache queue, performs overflow processing on the data cache queue, and reduces jitter in the network.

[0038] The optimal queue length at the previous moment, the change rate of the marking intensity, and the preset standard queue length are comprehensively considered to output the optimal queue length;

[0039] A portion of the data cache queue having a burst traffic label and an overload traffic label that is larger than the optimal queue length is marked as queue overflow and is removed from the data cache queue.

[0040] In a second aspect, the present invention provides a data transmission method used in a computer, comprising the following steps:

[0041] Step 1: Burst traffic monitoring;

[0042] The number of data packets or data segments waiting to be processed or transmitted in each data cache queue is obtained at each preset time interval, and recorded as the queue length of the data cache queue. The difference of the queue length is calculated at each preset time interval, and recorded as the queue length gradient.

[0043] As a preferred embodiment of the present invention, the flow status is detected according to the queue length and the queue length gradient, and the flow status label corresponding to each data cache queue is output, specifically:

[0044] If the queue length of a data cache queue is less than the minimum threshold and the queue length gradient is less than the minimum threshold, the low traffic label of the data cache queue is output;

[0045] If the queue length of a data cache queue is greater than the maximum threshold and the queue length gradient is less than the minimum threshold, a high traffic label of the data cache queue is output;

[0046] If the queue length of a data cache queue is less than the maximum threshold and greater than the minimum threshold and the queue length gradient is greater than the maximum threshold, then the burst traffic label of the data cache queue is output;

[0047] If the queue length of a data cache queue is greater than the maximum threshold and the queue length gradient is greater than the maximum threshold, the overload traffic label of the data cache queue is output;

[0048] In other cases, the stable flow label of the output data buffer queue.

[0049] Each traffic condition label corresponds to a preset label strength and label position. The label position points to a specific position in the data cache queue, and the specific position corresponds to a data packet in the data cache queue.

[0050] Step 2: Dynamic editing position and mark intensity adjustment;

[0051] Set the mark position of the low-traffic tag to always be at the front of the data cache queue.

[0052] Set the tag position of the high traffic tag to always be at the back end of the data cache queue.

[0053] Set the burst traffic tag position to always be in the middle of the data cache queue.

[0054] Set the mark position of the overload traffic label to always be at the front of the data cache queue.

[0055] No additional adjustment is made to the marking position of the stable flow label.

[0056] The marking strength adjustment comprehensively considers the overall fluctuation degree of the historical queue length of the data cache queue, the single fluctuation degree and the time difference from the single fluctuation moment to the current time; comprehensively considers the overall fluctuation degree of the historical queue length gradient, the single fluctuation degree and the time difference from the single fluctuation moment to the current time.

[0057] Specifically, the overall fluctuation degree is the variance, and the single fluctuation degree is the difference between the queue length and the queue length gradient and their own average values.

[0058] As a preferred embodiment of the present invention, the labeling intensity is allowed to decay over time.

[0059] Step 3: Transmission flow control;

[0060] Control the data transmission rate and manage the data cache queue. Adjust the data transmission rate according to different network conditions to ensure maximum throughput and avoid network congestion and resource waste.

[0061] The data transmission rate control comprehensively considers the transmission rate adjustment base of each data cache queue, and the adjustment amount is inversely proportional to the sum of the marking strengths of each data cache queue and inversely proportional to the network transmission delay at the current moment.

[0062] The data cache queue management is related to the type of traffic condition label.

[0063] Specifically, for the data cache queue with high traffic labels, its transmission rate is increased and the transmission window size is adjusted;

[0064] For the data cache queue of low-traffic tags, the transmission rate is kept unchanged to avoid unnecessary waste of resources;

[0065] For the data buffer queue with burst traffic labels, adjust the congestion window.

[0066] Step 4: Data cache queue management;

[0067] Adjust the maximum length of the data cache queue and perform overflow processing to reduce jitter in the network.

[0068] The optimal queue length at the previous moment, the change rate of the marking intensity, and the preset standard queue length are comprehensively considered to output the optimal queue length;

[0069] A portion of the data cache queue having a burst traffic label and an overload traffic label that is larger than the optimal queue length is marked as queue overflow and is removed from the data cache queue.

[0070] Compared with the prior art, the present invention has the following beneficial effects:

[0071] 1. The present invention establishes a burst traffic warning mechanism based on the data cache queue status, and combines the real-time monitoring of the queue length gradient and traffic situation labels to more accurately determine the burst traffic, overload traffic and stable traffic in the network. This precise traffic detection can effectively prevent network congestion and resource waste, while ensuring the efficiency and fairness of data transmission;

[0072] 2. The dynamic marking module of the present invention adjusts the marking position and marking strength according to the flow situation label, and reasonably controls the flow adjustment position and strength of the data packet. Through accurate marking adjustment, it helps to carry out timely flow control in the network, avoid premature or late suspension of data packet transmission, improve the stability and throughput of data transmission, and avoid resource waste and overload;

[0073] 3. The present invention solves the queue oscillation phenomenon existing in traditional networks and improves the stability of queues by optimizing the maximum length of queues, overflow processing and marking mechanisms. When encountering burst traffic or overload traffic, the optimal queue length is calculated and queue overflow management is dynamically adjusted to effectively reduce network fluctuations and improve the overall network stability and transmission efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings:

[0075] Figure 1 is a system block diagram of the present invention;

[0076] Figure 2 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0077] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0078] See also Figure 1 As shown, a data transmission system used in a computer includes a burst flow monitoring module, a dynamic marking module, a transmission control module and a queue management module.

[0079] The burst traffic monitoring module establishes a burst traffic warning mechanism based on the data queue status. The burst traffic warning mechanism dynamically monitors the changes in the data cache queue, calculates the queue length gradient to determine whether burst traffic occurs, and outputs the corresponding traffic situation label. The specific process is as follows:

[0080] The number of data packets or data segments waiting to be processed or transmitted in each data cache queue i is obtained at each preset time interval t, and is recorded as the data queue length Qi(t). The queue length gradient ΔQi(t) is calculated by the formula ΔQi(t)=Qi(t)-Qi(t-1) at each preset time interval t. Among them, Qi(t-1) is the queue length of the previous preset time interval.

[0081] Wherein, i is the number of the data cache queue, i=1, 2, ..., n; and n is the total number of cache queues.

[0082] If the queue length gradient ΔQi(t) is greater than a preset threshold ΔQiMax, it is determined that burst traffic has occurred in the cache queue i.

[0083] Furthermore, the traffic status is detected according to the queue length Qi(t) and the queue length gradient ΔQi(t), and the traffic status label corresponding to each data cache queue i is output, which is specifically:

[0084] If the queue length Qi(t) is less than the minimum threshold QiMin and the queue length gradient ΔQi(t) is less than the minimum threshold ΔQiMin, then the low flow label Li of the output data cache queue i is Low;

[0085] If the queue length Qi(t) is greater than the maximum threshold QiMax and the queue length gradient ΔQi(t) is less than the minimum threshold ΔQiMin, the high flow label Li of the output data cache queue i is High;

[0086] If the queue length Qi(t) is less than the maximum threshold QiMax and greater than the minimum threshold QiMin and the queue length gradient ΔQi(t) is greater than the maximum threshold Max, then the burst traffic label Li of the output data cache queue i is Burst;

[0087] If the queue length Qi(t) is greater than the maximum threshold QiMax and the queue length gradient ΔQi(t) is greater than the maximum threshold Max, then the overload traffic label Li of the output data cache queue i is set to Overload;

[0088] In other cases, the stable traffic label Li of the output data cache queue i is equal to Stable.

[0089] Each traffic condition label corresponds to a preset label strength Strength (Li, t) and a label position. The label position points to a specific position in the data cache queue, and the specific position corresponds to a data packet in the data cache queue.

[0090] It should be noted that the marking position and marking strength are crucial for adjusting the flow control strategy. The marking position determines where to start flow control, affecting the timeliness and fairness of transmission; the marking strength determines the flow control strength applied at the marking position, affecting the strength of flow restriction during data transmission. The marking position and marking strength jointly affect the network performance of each data cache queue during transmission, such as throughput, latency, and bandwidth utilization.

[0091] The queue length gradient calculation and flow status detection are performed on all number symbols i at preset time intervals t to obtain flow status labels of all data cache queues i=1, 2, . . . , n.

[0092] The dynamic marking module identifies the traffic condition label identified by the burst traffic monitoring module and selects a corresponding marking adjustment strategy.

[0093] The specific process of adjusting the mark position is as follows:

[0094] Set the mark position of the low-traffic tag to always be at the front of the data cache queue.

[0095] It should be noted that when the traffic is small, early marking helps to prevent data cache overflow in time and avoid waste of resources.

[0096] Set the tag position of the high traffic tag to always be at the back end of the data cache queue.

[0097] It should be noted that when the traffic is large, delayed marking can avoid premature transmission suspension caused by premature marking.

[0098] Set the burst traffic tag position to always be in the middle of the data cache queue.

[0099] It should be noted that the data cache queue corresponding to the burst traffic label has significant changes in queue length and gradient. The intermediate position mark can effectively cope with the short-term impact of burst traffic, perform congestion control as early as possible, and avoid the data cache queue being too long.

[0100] Set the mark position of the overload traffic label to always be at the front of the data cache queue.

[0101] It should be noted that overloaded traffic means that network resources are close to saturation. Early marking can help reduce the load of subsequent data flows and avoid congestion as much as possible.

[0102] No additional adjustment is made to the marking position of the stable flow label.

[0103] It should be noted that, without additional adjustment, as data packets enter and exit the data cache queue, the traffic status label attached to the data packet will be removed from the data cache queue as the data packet is transferred.

[0104] The specific process of adjusting the labeling intensity is as follows:

[0105] Obtain the historical queue length Qi(t) of the data cache queue i, where t=1, 2, ..., T; obtain the historical queue length gradient ΔQi(t) of the data cache queue i, where t=1, 2, ..., T; where T is the current time.

[0106] By formula Calculate the mean μ(Qi) of the historical queue length Qi(t) and the mean μ(ΔQi) of the historical queue length gradient ΔQi(t);

[0107] By formula Calculate the variance σ(Qi) of the historical queue length Qi(t) and the mean σ(ΔQi) of the historical queue length gradient ΔQi(t);

[0108] Furthermore, through the formula Calculate the rolling volatility V of the historical queue length Qi(t) Qi and the rolling volatility V of the historical queue length gradient ΔQi(t) ΔQi α1 and α2 are preset impact factors, β is a preset impact index, and k1 and k2 are preset weight coefficients.

[0109] Through the formula Strength(Li, t) = Strength(Li, t) × e -C×(T-t) +Gain(Li, t)

[0110] The labeling intensity is dynamically adjusted to obtain the adjusted final labeling intensity Strength(Li, t), where C is the preset decay rate.

[0111] The transmission control module controls the data transmission rate and data cache queue management according to the traffic model and network status. By adjusting the data transmission rate according to different network conditions, the throughput is maximized while avoiding network congestion and resource waste. The specific process is as follows:

[0112] By formula Calculate the transmission rate adjustment base R of each data cache queue i i, where Bi is the maximum allowable bandwidth of data cache queue i. Where Delay is the network transmission delay, δ is the preset weight factor, and Delay(t) is the network transmission delay at time t.

[0113] Furthermore, the transmission rate and congestion window are adjusted based on the traffic situation label to control the congestion in data transmission and ensure the stability and efficiency of each flow in the network. The specific process is as follows:

[0114] For the data cache queue with high traffic labels, increase its transmission rate by a transmission rate adjustment base R i , and adjust the transmission window size;

[0115] For the data cache queue of low-traffic tags, the transmission rate is kept unchanged to avoid unnecessary waste of resources;

[0116] For the data cache queue with burst traffic labels, adjust the congestion window to allow for short-term traffic growth.

[0117] The congestion window adjustment rule is: i (t) = W i (t-1)+η×[1-ECN(i,t)]where W i (t-1) is the congestion window of the data cache queue i at the previous moment; wherein ECN(i, t) is the specific value of the preset explicit congestion notification adjustment tag.

[0118] It should be noted that the congestion window is a flow control mechanism in the network, usually used in the TCP protocol, to limit the maximum amount of data that the sender can send without receiving confirmation. Its function is to avoid data packet loss and overload caused by network congestion.

[0119] The queue management module optimizes the queue management strategy to deal with the queue oscillation phenomenon in traditional networks. It reduces the jitter in the network and improves the stability of the data cache queue by adjusting the maximum length of the data cache queue and the overflow processing of the data cache queue.

[0120] For a data cache queue i with a burst traffic label and an overload traffic label, the formula Calculate the optimal queue length Qi(t) Max Qi(t-1) Max is the optimal queue length at the previous moment; Qi(t-1) is the queue length at the previous moment; strength(Li, t-1) is the marking strength at the previous moment; Q0 is the preset standard queue length;

[0121] The data with burst traffic label and overload traffic label are cached in queue i which is larger than the optimal queue length Qi(t).Max The part is marked as queue overflow and removed from the data cache queue.

[0122] The above process is performed every preset time interval t on all data cache queues i.

[0123] See also Figure 2 As shown, a data transmission method used in a computer comprises the following steps:

[0124] Step 1: Burst traffic monitoring;

[0125] The number of data packets or data segments waiting to be processed or transmitted in each data cache queue i is obtained at each preset time interval t, and is recorded as the data queue length Qi(t). The queue length gradient ΔQi(t) is calculated by the formula ΔQi(t)=Qi(t)-Qi(t-1) at each preset time interval t. Among them, Qi(t-1) is the queue length of the previous preset time interval.

[0126] Where i is the number of the data cache queue, i=1, 2, . . . , n; where n is the total number of data cache queues.

[0127] If the queue length gradient ΔQi(t) is greater than a preset threshold ΔQiMax, it is determined that burst traffic has occurred in the data cache queue i.

[0128] Furthermore, the traffic status is detected according to the queue length Qi(t) and the queue length gradient ΔQi(t), and the traffic status label corresponding to each data cache queue i is output, which is specifically:

[0129] If the queue length Qi(t) is less than the minimum threshold QiMin and the queue length gradient ΔQi(t) is less than the minimum threshold ΔQiMin, then the low flow label Li of the output data cache queue i is Low;

[0130] If the queue length Qi(t) is greater than the maximum threshold QiMax and the queue length gradient ΔQi(t) is less than the minimum threshold ΔQiMin, the high flow label Li of the output data cache queue i is High;

[0131] If the queue length Qi(t) is less than the maximum threshold QiMax and greater than the minimum threshold QiMin and the queue length gradient ΔQi(t) is greater than the maximum threshold Max, then the burst traffic label Li of the output data cache queue i is Burst;

[0132] If the queue length Qi(t) is greater than the maximum threshold QiMax and the queue length gradient ΔQi(t) is greater than the maximum threshold Max, then the overload traffic label Li of the output data cache queue i is set to Overload;

[0133] In other cases, the stable traffic label Li of the output data cache queue i is equal to Stable.

[0134] Each traffic condition label corresponds to a preset label strength Strength (Li, t) and a label position. The label position points to a specific position in the data cache queue, and the specific position corresponds to a data packet in the data cache queue.

[0135] The queue length gradient calculation and flow status detection are performed on all number symbols i at preset time intervals t to obtain flow status labels of all data cache queues i=1, 2, . . . , n.

[0136] Step 2: Dynamic editing position and mark intensity adjustment;

[0137] Set the mark position of the low-traffic tag to always be at the front of the data cache queue.

[0138] Set the tag position of the high traffic tag to always be at the back end of the data cache queue.

[0139] Set the burst traffic tag position to always be in the middle of the data cache queue.

[0140] Set the mark position of the overload traffic label to always be at the front of the data cache queue.

[0141] No additional adjustment is made to the marking position of the stable flow label.

[0142] The specific process of adjusting the labeling intensity is as follows:

[0143] Obtain the historical queue length Qi(t) of the data cache queue i, where t=1, 2, ..., T; obtain the historical queue length gradient ΔQi(t) of the cache queue i, where t=1, 2, ..., T; where T is the current time.

[0144] By formula Calculate the mean μ(Qi) of the historical queue length Qi(t) and the mean μ(ΔQi) of the historical queue length gradient ΔQi(t);

[0145] By formula Calculate the variance σ(Qi) of the historical queue length Qi(t) and the mean σ(ΔQi) of the historical queue length gradient ΔQi(t);

[0146] Furthermore, through the formula Calculate the rolling volatility V of the historical queue length Qi(t) Qi and the rolling volatility V of the historical queue length gradient ΔQi(t) ΔQi.

[0147] α1 and α2 are preset influencing factors, β is a preset influencing index, and k1 and k2 are preset weight coefficients.

[0148] Through the formula Strength(Li, t) = Strength(Li, t) × e -C×(T-t) +Gain(Li, t)

[0149] The labeling intensity is dynamically adjusted to obtain the adjusted final labeling intensity Strength(Li, t), where C is the preset decay rate.

[0150] Step 3: Traffic transmission control;

[0151] By formula Calculate the transmission rate adjustment base R of each data cache queue i i , where Bi is the maximum allowed bandwidth of data cache queue i. Where Delay is the network transmission delay, where δ is the preset weight factor, and where Delay(t) is the network transmission delay at time t.

[0152] Furthermore, the transmission rate and congestion window are adjusted based on the traffic situation label to control the congestion in data transmission and ensure the stability and efficiency of each flow in the network. The specific process is as follows:

[0153] For the data cache queue with high traffic labels, increase its transmission rate by a transmission rate adjustment base R i , and adjust the transmission window size;

[0154] For the data cache queue of low-traffic tags, the transmission rate is kept unchanged to avoid unnecessary waste of resources;

[0155] For the data cache queue with burst traffic labels, adjust the congestion window to allow for short-term traffic growth.

[0156] The congestion window adjustment rule is: i (t) = W i (t-1)+η×[1-ECN(i,t)]where W i (t-1) is the congestion window of the data cache queue i at the previous moment; wherein ECN(i, t) is the specific value of the preset explicit congestion notification adjustment tag.

[0157] Step 4: Data cache queue management;

[0158] For a data cache queue i with a burst traffic label and an overload traffic label, the formula Calculate the optimal queue length Qi(t) MaxQi(t-1) Max is the optimal queue length at the previous moment; Qi(t-1) is the queue length at the previous moment; strength(Li, t-1) is the marking strength at the previous moment; Q0 is the preset standard queue length;

[0159] The data with burst traffic label and overload traffic label are cached in queue i which is larger than the optimal queue length Qi(t). Max The part is marked as queue overflow and removed from the data cache queue.

[0160] The above process is performed every preset time interval t on all data cache queues i.

[0161] It should be understood that the terms “include” and “comprising” used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0162] It should also be understood that the terms used in this disclosure are only for the purpose of describing specific embodiments and are not intended to limit the disclosure. As used in this disclosure and the claims, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in this disclosure and the claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations;

[0163] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A data transmission system used in a computer, comprising a burst flow monitoring module and a dynamic marking module, characterized in that: The burst traffic monitoring module establishes a burst traffic early warning mechanism based on the data cache queue status; The burst traffic warning mechanism dynamically monitors the changes in the data cache queue and outputs corresponding traffic situation labels, including low traffic labels, high traffic labels, burst traffic labels, overload traffic labels and stable traffic labels; The dynamic marking module identifies the traffic condition label identified by the burst traffic monitoring module, and selects a corresponding marking adjustment strategy, including marking position adjustment and marking strength adjustment.

2. A data transmission system for use in a computer according to claim 1, further comprising a transmission control module and a queue management module, characterized in that: The transmission control module controls the data transmission rate and performs queue management; it adjusts the data transmission rate according to different network conditions to ensure maximum throughput; The queue management module adjusts the maximum length of the data cache queue and performs overflow processing on the data cache queue; The optimal queue length at the previous moment, the change rate of the marking intensity, and the preset standard queue length are comprehensively considered to output the optimal queue length; A portion of the data cache queue having a burst traffic label and an overload traffic label that is larger than the optimal queue length is marked as queue overflow and is removed from the data cache queue.

3. A data transmission system for use in a computer according to claim 1, characterized in that: The dynamic monitoring is achieved by analyzing the queue length and the queue length gradient; The queue length refers to the number of data packets or data segments waiting to be processed or transmitted in each data cache queue; The queue length gradient refers to the change in the queue length in each data cache queue over time.

4. A data transmission system for use in a computer according to claim 1, characterized in that: The traffic status is detected according to the queue length and the queue length gradient, and the corresponding relationship of the traffic status labels corresponding to each data cache queue is output as follows: If the queue length of a data cache queue is less than the minimum threshold and the queue length gradient is less than the minimum threshold, the low traffic label of the data cache queue is output; If the queue length of a data cache queue is greater than the maximum threshold and the queue length gradient is less than the minimum threshold, a high traffic label of the data cache queue is output; If the queue length of a data cache queue is less than the maximum threshold and greater than the minimum threshold and the queue length gradient is greater than the maximum threshold, then the burst traffic label of the data cache queue is output; If the queue length of a data cache queue is greater than the maximum threshold and the queue length gradient is greater than the maximum threshold, the overload traffic label of the data cache queue is output; Otherwise, the stable traffic label of the output buffer queue.

5. A data transmission system for use in a computer according to claim 4, characterized in that: Each traffic condition label corresponds to a preset label strength and label position; the label position points to a specific position in the data cache queue, and the specific position corresponds to a data packet in the data cache queue.

6. A data transmission system for use in a computer according to claim 1, characterized in that: The specific process of adjusting the mark position is as follows: Set the mark position of the low-traffic tag to always be at the front of the data cache queue; Set the tag position of the high-traffic tag to always be at the back end of the data cache queue; Set the mark position of the burst traffic label to always be in the middle of the data cache queue; The mark position of the overload traffic label is set to always be at the front of the data cache queue; No additional adjustment is made to the marking position of the stable flow label.

7. A data transmission system for use in a computer according to claim 1, characterized in that: The marking intensity adjustment comprehensively considers the overall fluctuation degree of the historical queue length of the data cache queue, the single fluctuation degree and the time difference from the single fluctuation moment to the current time; comprehensively considers the overall fluctuation degree of the historical queue length gradient, the single fluctuation degree and the time difference from the single fluctuation moment to the current time; The overall fluctuation degree is the variance, and the single fluctuation degree is the difference between the queue length and the queue length gradient and their own average value.

8. A data transmission system for use in a computer according to claim 1, characterized in that: Allows the marker intensity to decay over time.

9. A data transmission system for use in a computer according to claim 2, characterized in that: The data transmission rate control comprehensively considers the transmission rate adjustment base of each data cache queue, and the adjustment amount is inversely proportional to the sum of the marking strengths of each data cache queue and inversely proportional to the network transmission delay at the current moment; The queue management described is related to the type of traffic situation label; Specifically: for the data cache queue with high traffic labels, increase its transmission rate and adjust the transmission window size; For the data cache queue with low traffic tags, the transmission rate is kept unchanged; For the data buffer queue with burst traffic labels, adjust the congestion window.

10. A data transmission method used in a computer, characterized in that A data transmission system for use in a computer as claimed in any one of claims 1 to 9, comprising the following steps: Step 1: Burst traffic monitoring; At every preset time interval, the number of data packets or data segments waiting to be processed or transmitted in each data cache queue is obtained, which is recorded as the queue length of the data cache queue; at every preset time interval, the difference of the queue length is calculated, which is recorded as the queue length gradient; The traffic status is detected based on the queue length and the queue length gradient, and the traffic status labels corresponding to each data cache queue are output, specifically: If the queue length of a data cache queue is less than the minimum threshold and the queue length gradient is less than the minimum threshold, the low traffic label of the data cache queue is output; If the queue length of a data cache queue is greater than the maximum threshold and the queue length gradient is less than the minimum threshold, a high traffic label of the data cache queue is output; If the queue length of a data cache queue is less than the maximum threshold and greater than the minimum threshold and the queue length gradient is greater than the maximum threshold, then the burst traffic label of the data cache queue is output; If the queue length of a data cache queue is greater than the maximum threshold and the queue length gradient is greater than the maximum threshold, the overload traffic label of the data cache queue is output; In other cases, the stable flow label of the output data buffer queue; Each traffic condition label corresponds to a preset label strength and label position; the label position points to a specific position in the data cache queue, and the specific position corresponds to a data packet in the data cache queue; Step 2: Dynamic editing position and mark intensity adjustment; Set the mark position of the low-traffic tag to always be at the front of the data cache queue; Set the mark position of the high-traffic tag to always be at the back end of the data cache queue; Set the burst traffic label position to always be in the middle of the data cache queue; The mark position of the overload flow label is set to always be at the front of the data cache queue; No additional adjustments are made to the marking position of the stable flow label; The marking intensity adjustment comprehensively considers the overall fluctuation degree of the historical queue length of the data cache queue, the single fluctuation degree and the time difference from the single fluctuation moment to the current time; comprehensively considers the overall fluctuation degree of the historical queue length gradient, the single fluctuation degree and the time difference from the single fluctuation moment to the current time; The overall fluctuation degree is the variance, and the single fluctuation degree is the difference between the queue length and the queue length gradient and its own average value; Let the marking intensity decay over time; Step 3: Transmission flow control; Control the data transmission rate and manage the data cache queue; adjust the data transmission rate according to different network conditions to ensure maximum throughput; The data transmission rate control comprehensively considers the transmission rate adjustment base of each data cache queue, and the adjustment amount is inversely proportional to the sum of the marking strengths of each data cache queue and inversely proportional to the network transmission delay at the current moment; The data cache queue management is related to the type of traffic condition label; For data cache queues with high traffic labels, increase their transmission rate and adjust the transmission window size; For the data cache queue of low-traffic tags, the transmission rate is kept unchanged to avoid unnecessary waste of resources; For the data cache queue with burst traffic labels, adjust the congestion window; Step 4: Data cache queue management; Adjust the maximum length of the data cache queue and perform overflow processing to reduce jitter in the network; The optimal queue length at the previous moment, the change rate of the marking intensity, and the preset standard queue length are comprehensively considered to output the optimal queue length; A portion of the data cache queue having a burst traffic label and an overload traffic label that is larger than the optimal queue length is marked as queue overflow and is removed from the data cache queue.

Images