KR20140139406A - Randomization of packet size - Google Patents

Abstract

A method of randomizing a size of a packet is provided. A randomization method includes: identifying a maximum segment size (MSS) defined for transmission and reception of a packet with a reception terminal; and randomizing a size of the packet to be less than the defined MSS.

Description

[0002] RANDOMIZATION OF PACKET SIZE [0003]

This disclosure relates to a packet size randomization technique, and more particularly to a packet size randomization technique for a transmitting terminal that sends and receives packets to and from a receiving terminal.

Within the data center, the use of many-to-one transmission patterns (eg, MapReduce) is increasing. In this case, the data flows compete with each other at the receiving ingress point. For low-cost switches that are widely used in the data center, drop-tail queuing policy is used, so packet drops occur fairly to input ports of ingress points. do.

 However, if the number of received data flows for a particular X input port (X) is less than the number of received data flows for another Y input port (Y), the data flows coming in to the X input port will have a Transmission Control Protocol (timeout), which can result in severe loss of TCP throughput. This is called the TCP outcast problem.

Accordingly, there is a need to develop a technique for solving the TCP outcast problem.

The present disclosure has been devised in response to the above-described technology development, and the present disclosure provides a packet size randomization technique of a transmitting terminal that randomizes a size of a packet and transmits and receives a packet to and from a receiving terminal.

According to an embodiment of the present invention, there is provided a random number generation method of a transmitting terminal, comprising: determining a maximum segment size (MSS) defined when transmitting and receiving a packet to and from a receiving terminal; and determining a size of the packet to be less than the defined maximum segment size And a step of marshalling the signal.

The step of randomizing the size of the packet may comprise the step of determining the degree of randomization.

The step of determining the degree of randomization may receive a parameter that adjusts the degree of randomization and determine the degree of randomization based on the received parameter.

The step of determining the degree of randomization may determine the parameter for adjusting the degree of randomization of the transmitting terminal.

The transmitting terminal may further include a step of collecting randomization-related information for determining the degree of randomization.

The randomization related information may be at least one of multiplier information and congestion window information.

In the step of determining the degree of randomization, the multiplier may be collected through a parameter of a TCP stack (TCP stactk) of an operating system kernel.

The step of determining the degree of randomization may further comprise determining whether or not to timeout the transmission and reception of the packet in response to a variation of at least one of the multiplier information and the congestion windowing information .

Wherein the step of determining the degree of randomization determines whether a timeout of transmission and reception of the packet has not occurred when the value of the multiplier is a first value or if the value of the multiplier is a second value It can be determined that a timeout of transmission and reception of the packet has occurred.

The step of determining the degree of randomization may determine that the timeout of transmission and reception of the packet occurs continuously as the value of at least one of the multiplier information and the congestion windowing information increases.

The step of determining the degree of randomization may determine the degree of randomization according to at least one of the multiplier information and the congestion windowing information.

The degree of randomization may include at least one of a maximum value of the randomized value, a minimum value of the randomized value, an expected value of the randomized value, and a standard deviation of the distribution of the randomized value.

The step of randomizing the size of the packet may adjust the size of the payload of the packet through a randomization function in a transport layer.

The transmitting terminal may further include a step of transmitting a packet whose size is adjusted in the payload size to a lower layer.

The step of randomizing the size of the payload of the packet may segment a segment larger than the maximum segment unit input from the upper layer when Transmit segmentation offload (TSO) is activated.

The randomization technique of the transmitting terminal may perform randomization according to a software defined randomization algorithm in at least one of a router and a switch.

According to another embodiment of the present invention, there is provided a random number generation method of a transmitting terminal, comprising: monitoring at least one of multiplier information and congestion windowing information of a TCP stack (TCP stactk) of an operating system kernel; Determining whether the size of a packet to be transmitted is randomized, and transmitting the packet to the receiving terminal.

The step of determining whether the size of the payload is randomized may determine that the randomization is not performed when at least one of the multiplier information and the congestion windowing information is a first value.

Wherein if the at least one of the multiplier information and the congestion windowing information is a second value, the randomization is performed, and the degree of randomization increases as at least one of the multiplier information and the congestion windowing information increases .

The transmitting terminal according to another embodiment may be configured to determine a maximum segment size (MSS) defined when transmitting and receiving a packet to and from a receiving terminal, and to randomize the packet size to less than the defined maximum segment size And a communication unit for transmitting the packet to the receiving terminal.

According to various embodiments, a packet size randomization technique of a transmitting terminal that transmits / receives a packet to / from a receiving terminal by randomizing the size of at least one of a header and a payload of the packet can be provided. In particular, the TCP outcast problem can be solved by sizing at least one of the header and payload of the packet to less than the maximum segment size (MSS).

In particular, various embodiments can be applied to various implementations such as a TCP stack of an operating system, an offload engine of a network interface card (NIC), a router / switch, and the like.

In addition, when applied to the TCP stack of the operating system, it can be mounted on a Windows-based / Unix-based server-class OS, and thus compatibility can be maximized. It is expected that the use of the large segment offload of the NIC will become indispensable when the 10 Gbps is generalized in the future due to the improvement of the transmission speed, and it may be used in the hardware market by being attached to the offload engine to which various embodiments are applied.

In addition, various embodiments may be actively used in traffic management by being installed in a software defined router / switch or the like.

1 is a conceptual diagram for explaining a topology to which various embodiments are applied.
2 is a conceptual diagram for explaining a flow process in an ingress switch environment composed of two input ports (X, Y) and one output port (Z) according to a comparative example.
3 is a graph showing the relationship between the throughput-transmission / reception terminal RTT.
4 is a conceptual diagram for explaining a applicable method according to a comparative example.
FIG. 5 is a flowchart illustrating a randomization technique of a transmitting terminal according to an embodiment.
FIG. 6 is a flowchart for explaining a method for determining the degree of randomization according to an embodiment.
7 is a block diagram of a transmitting node according to one embodiment.
8A is a conceptual diagram of a packet that has been randomized according to an embodiment.
8B is a conceptual diagram for explaining a flow processing improvement according to an embodiment of the randomization technique.
9 shows experimental results on a network simulator.
10A is a graph of a congestion window in the example of CUBIC, which is a congestion control algorithm used by default in the Linux kernel.
10B is a graph showing a change in the degree of randomization of the transmitting terminal corresponding to the congestion window situation of FIG. 10A.

The present disclosure is capable of various modifications and various embodiments, and specific embodiments are illustrated and described in the drawings. It should be understood, however, that this disclosure is not intended to be limited to any particular embodiment, but is intended to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the disclosure.

The terms including ordinal, such as second, first, etc., may be used to describe various elements, but the elements are not limited to these terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present disclosure, the second component may be referred to as a first component, and similarly, the first component may also be referred to as a second component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the disclosure. The singular expressions include plural expressions unless the context clearly dictates otherwise. In the present application, the terms "comprises" or "having" and the like are used to specify that there is a feature, a number, a step, an operation, an element, a component or a combination thereof described in the specification, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs. Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are to be interpreted as either ideal or overly formal in the sense of the present application Do not.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings, wherein like or corresponding elements are denoted by the same reference numerals, and redundant description thereof will be omitted.

1 is a conceptual diagram for explaining a topology to which various embodiments are applied.

1, the topology to which various embodiments are applied includes at least one Core (Core 1 to Core 4), at least one Aggr (Aggr 1 to Aggr 8), at least one Edge (Edge 1 to Edge 8) , And at least one terminal (R, S1 to S15).

At least one Aggr (Aggr1 to Aggr8) may be connected to each of at least one Core (Core 1 to Core 4). In addition, one edge (Edge 1 to Edge 8) may be connected to each of at least one Aggr (Aggr1 to Aggr8). At least one terminal (R, S1 to S15) may be connected to each of at least one edge (Edge 1 to Edge 8). Core, Aggr, Edge, and terminal can form a fat-tree structure in the above-described order. That is, the topology according to the embodiment of FIG. 1 may be based on a fat-tree topology and may be referred to as a multi-rooted and hierachical topology. On the other hand, the above-described topology is merely exemplary, and one skilled in the art will readily understand that the embodiments of the present disclosure can be applied to various topologies and that there is no limitation on the type of topology applied.

Core (Core 1 to Core 4), Aggr (Aggr 1 to Aggr 8), and Edge (Edge 1 to Edge 8) in FIG. 1 may be a switch set constituting the topology. On the other hand, the terminals R, S1 to S15 may be leaf nodes of a topology in which data flows or packets are transmitted and received.

On the other hand, in a many-to-one transmission application such as MapReduce according to the comparative example, multiple flows are concentrated in each port of the ingress switch of the receiving terminal.

Eventually, there is a competition between the flows to enter the ingress queue of the ingress switch. The commercial switch according to the comparative example employs a drop-tail FIFO queuing method. In the case of the comparative example, a port blackout problem may occur. Here, the port blackout refers to a phenomenon in which packets arriving at an arbitrary port of the ingress switch are continuously dropped.

2 is a conceptual diagram for explaining a flow process in an ingress switch environment composed of two input ports (X, Y) and one output port (Z) according to a comparative example.

The input time of the first packet (X1, X2, X3, X4) input to the X input port (Port X) and the input of the second packet (Y1, Y2, Y3, Y4) input to the Y input port The arrival times of the first packets (X1, X2, X3, X4) and the arrival times of the second packets (Y1, Y2, Y3, Y4) may be different. For example, in the comparative example of Fig. 2, the arrival time (T [X1], T [X2], T [X3], T [X4]) of the first packet (X1, X2, X3, X4) The arrival times T [Y1], T [Y2], T [Y3], and T [Y4] of the Y1, Y2, Y3 and Y4 may be faster. Since the port according to the comparative example employs a drop-tail FIFO queuing scheme, the first packet (X1, X2, X3, X3) input to the input port (Port X) X4 are all dropped. As a result, only the second packets (Y1, Y2, Y3, Y4) input to the Y input port (Port Y) are output to the output port (Z).

Port blackout occurs not on a specific port but on all ports. However, a continuous packet drop causes a TCP time-out, which results in a problem of rapidly lowering the sending rate. This problem is called the TCP outcast problem. TCP outcast problems are particularly likely to occur in tiered topologies such as fat-trees. For example, in the case of many-to-one (15-to-1) between one receiving terminal R and fifteen transmitting terminals S in Fig. 1, There are three types of flow as follows.

1) 2hop flow: S1-> R

2) 4hop flows: S2-> R, S3-> R

3) 6hop flows: [S4 ,,, S15] -> R

In general, TCP throughput is inversely proportional to RTT between transmitting and receiving terminals. Therefore, it can be expected that the throughput is large in order of 1)> 2)> 3), but the result is determined as shown in FIG. 3 is a graph showing the relationship between the throughput-transmission / reception terminal RTT. The result of FIG. 3 relates to the port blackout occurring in Edge 1 of FIG.

Edge1 uses two input ports, Edge1: S1 and Edge1: Aggr1. Here, Edge1: Aggr2 can be assumed to be redundant. At this time, the number of flows entering Edge1: S1 is one (S1-> R). On the other hand, the number of flows to Edge1: Aggr1 is 14. In this case, port blackout occurs in Edge1: S1 and Edge1: Aggr1 respectively. In the case of Edge1: S1, since all packets dropped continuously by port blackout are part of S1-> R, S1-> R causes a TCP timeout by congestion control of the TCP, Lower the congestion window size to 1. That is, a serious transmission speed degradation may be caused.

On the other hand, Edge1: Aggr1 contains 14 flows. Therefore, even if a packet is continuously dropped, the loss may not be large in each flow.

A method according to a comparative example for solving this problem was tested as shown in Fig. 4 is a conceptual diagram for explaining a applicable method according to a comparative example. RED / SFQ does not perform drop-tail queuing as a link layer solution. As a result, the outcast problem does not occur.

However, RED equipment is expensive and exhibits RTT bias. SFQ is very expensive and rarely used in commercial switches. TCP pacing is a transport layer solution that adjusts the time interval between packets during packet transmission. However, in order to apply the above problem, microsecond level time accuracy must be guaranteed. However, this can not be implemented in the current CPU / OS architecture. Equal-length routing is a network layer solution that diverts all flows to the same path, but instead deliberately chooses a detour path instead of a shortest path, There is a problem of seriously degrading the efficiency.

The embodiment of the present disclosure proposes a randomization technique for randomizing a packet size, and more specifically, a packet payload size, in order to solve the problem of the above-described comparative example.

In the conventional packet size determination method according to the comparative example, for example, the payload size of the TCP payload is set to be the maximum segment size (MSS), which is always equal to the maximum value. In the conventional packet size determination method according to the comparative example, the above-described payload size determination method is used in order to reduce the header overhead. However, in order to solve the problem of the comparative example in the case where a TCP outcast problem occurs, the embodiment of the present disclosure provides a method of randomizing the TCP payload size within a range smaller than the MSS .

FIG. 5 is a flowchart illustrating a randomization technique of a transmitting terminal according to an embodiment.

In step 510, the transmitting terminal can determine the maximum segment size (MSS) in transmitting and receiving packets with the receiving terminal. Here, the maximum segment size may mean the maximum allowable size in the course of performing packet segmentation.

The transmitting terminal and the receiving terminal can determine the size of the maximum segment in the process of installing the predetermined protocol. Or the transmitting terminal and the receiving terminal may exchange messages with each other to determine the size of the maximum segment. In addition, the transmitting terminal may receive information on the maximum segment size from the manager.

At step 520, the transmitting terminal may randomize the packet size to less than the predetermined maximum segment size. For example, the sending terminal can randomize the size of the payload of a packet to less than a predetermined maximum segment size.

The transmitting terminal can determine the size of the payload as a value between [1, MSS]. The sending terminal can randomize the packet size to less than the maximum segment size since randomization above the maximum segment size can cause separate fragmentation.

Meanwhile, the transmitting terminal may determine the degree of randomness in randomizing the packet payload size. Here, the degree of randomization may be at least one of the maximum value (randMAX) of the randomized value, the lowest value (randMIN) of the randomized value, the expected value (randEXP) of the randomized value, and the standard deviation randSTDV of the randomized value distribution One can be included. The transmitting terminal can adjust the degree of randomization by parameters.

In one embodiment, the transmitting terminal may operate in a passive mode to determine the degree of randomization. For example, the transmitting terminal may set the degree of randomization according to the input or preset operator's policy. Many TCP parameters are set to default values in the operating system and can be changed by a system administrator or an operator when necessary, so that the transmitting terminal can set up and change the degree of randomization.

As an example, in the case of a data center, a system administrator, a program, or a specific algorithm can grasp the characteristics of a particular many-to-one flow. A system administrator, a program, or a specific algorithm can analyze the degree of specialized randomization according to various cases in advance, and then the degree of randomization can be set manually based on the analysis result.

In another embodiment, the transmitting terminal may operate in an automatic mode to determine the degree of randomization. The transmitting terminal can collect various randomization related information for determining the degree of randomization. The transmitting terminal can determine the degree of randomization based on the collected randomization-related information. A transmitting terminal according to an exemplary embodiment may collect a TCP multiplier as randomization related information, and may monitor a TCP multiplier, for example.

Experimental results show that the multiplier used for backoff in TCP retransmit timeout is an important factor in predicting TCP outcast problem.

More specifically, if the transmitting terminal does not receive an ack for a packet to be transmitted, it can enter a timeout. That is, the transmitting terminal can not transmit a packet for a specific time. On the other hand, when the timeout is terminated, the transmitting terminal may not be able to resume transmission by slow start. In this case, if no acknowledgment is received for the transmitted packet, the timeout is restarted. As described above, because the timeouts occur consecutively, the timeout period increases exponentially. If such a situation is repeated, the transmission efficiency is lowered.

In one embodiment, the transmitting terminal may monitor the fluctuation of the multiplier of a particular flow. The transmitting terminal can easily grasp by monitoring the multiplier variation whether or not the flow enters the timeout and the transmission environment deteriorates.

In one embodiment, the sending terminal can grasp the multiplier of a particular flow through the parameters of the operating system's kernel TCP stack. For example, the transmitting terminal may determine that no timeout occurs when the multiplier is a first number, e.g., 1. In addition, the transmitting terminal can determine that the multiplier increases when a timeout occurs. Alternatively, the transmitting terminal may determine that a timeout occurs consecutively when the multiplier increases. For example, if the multiplier is 2, the timeout occurs once. When the multiplier is 4, the timeout occurs twice. When the multiplier is 8, the timeout occurs 3 times It can be judged that it has occurred. The transmitting terminal can determine the linear relationship between the frequency of occurrence of the timeout and the multiplier described above when the product of the RTO and the multiplier is smaller than the maximum value of the timeout.

FIG. 6 is a flowchart for explaining a method for determining the degree of randomization according to an embodiment.

In step 610, the sending terminal may monitor the multiplier. As described above, the transmitting terminal can grasp the multiplier of the specific flow through the parameters of the kernel TCP stack of the operating system.

In step 620, the transmitting terminal may enter step 630 and not perform randomization if the multiplier is a first number, e.g., 1. That is, the transmitting terminal can determine that no timeout has occurred and can set the degree of randomization to zero.

On the other hand, in step 640, the transmitting terminal can determine the degree of randomization corresponding to the multiplier degree. For example, if the product of the RTO and the multiplier is smaller than the maximum value of the timeout, the transmitting terminal can increase the degree of randomization as the multiplier increases. That is, the transmitting terminal determines that the frequency of the timeout increases as the multiplier increases, and thus the degree of randomization can be increased. It will be readily understood by those skilled in the art that the various embodiments do not limit the increase in the degree of randomization and that the range of the right is not limited by the increase in the degree of randomization.

The randomization function according to various embodiments may be implemented according to various embodiments as follows.

The randomization function according to one embodiment can be implemented in the TCP stack of the operating system kernel. In more detail, the transmitting module may adjust the payload size through the randomization function in the transport layer, and then transmit the packet to the lower layer.

In this embodiment, since the transport layer itself responsible for congestion control carries out up to the segment size randomization, it is possible to implement randomization based on various information. In addition, it is possible to implement the randomization according to the embodiment only by modifying the operating system kernel (Software).

The randomization function according to another embodiment may be implemented in an offload engine of a network interface card (NIC). In particular, when Transmit segmentation offload (TSO) is enabled, the NIC can receive a large segment larger than the MSS, rather than an MSS segment, from the upper layer. The offload engine of the NIC can segment the input large segment. Accordingly, the consumption of the CPU cycle can be minimized and a hardware-level packet size randomization can be realized. Meanwhile, in the transport layer, various information necessary for packet size randomization from the NIC, for example, the randomization module and the input parameters of the randomization degree calculation module, can be received.

The randomization function according to another embodiment may be implemented in a software defined router / switch, in which case the application of the randomization algorithm may be possible.

7 is a block diagram of a transmitting node according to one embodiment.

The transmitting node 700 may include a control unit 710 and a communication unit 720.

The control unit 710 can generate a packet to be transmitted by the communication unit 720. [ The control unit 710 can generate a packet according to a predetermined protocol and determine the packet size. The control unit 710 can determine the predetermined maximum segment size when transmitting and receiving packets with the receiving terminal. The control unit 710 may randomize the size of the payload of the packet to less than the predetermined maximum segment size. The controller 710 may be implemented as an IC chip, a microprocessor, a mini computer, or the like.

The communication unit 720 can transmit the packet whose size is unordered to the receiving terminal. The communication unit 720 may include various communication modules such as an antenna, a comb / modulator, a frequency processing unit, and a filter unit.

The control unit 710 may determine the degree of randomization. The controller 710 may operate in a passive mode in which the parameter for adjusting the degree of randomization is received from outside and the degree of randomization is determined. Alternatively, the control unit 710 may operate in an automatic mode in which the transmitting terminal determines parameters for adjusting the degree of randomization.

The control unit 710 may collect the randomization-related information for determining the degree of randomization. Here, the randomization related information may be multiplier information as described above. The controller 710 may collect the multiplier through a parameter of the TCP stack of the operating system kernel.

The control unit 710 can determine whether or not to time out the transmission / reception of the packet in response to the variation of the multiplier. If the value of the multiplier is a first value, the control unit 710 determines that the timeout of the transmission / reception of the packet has not occurred, or if the value of the multiplier is the second value, Out can be determined to have occurred. The control unit 710 may determine that the timeout of transmission and reception of the packet occurs continuously as the value of the multiplier increases.

The control unit 710 can determine the degree of randomization according to the numerical value of the multiplier. The degree of randomization may include at least one of a maximum value of the randomized value, a minimum value of the randomized value, an expected value of the randomized value, and a standard deviation of the distribution of the randomized value.

The controller 710 according to an exemplary embodiment of the present invention may randomize the payload size of the packet by adjusting a size of the payload through a randomization function in a transport layer. The communication unit 720 may transmit the packet whose size is adjusted to the payload size to the lower layer.

When the TSO (Transmit segmentation offload) is activated, the controller 710 may segment a segment larger than the maximum segment unit input from the upper layer.

The controller 710 may perform randomization according to a software defined randomization algorithm in at least one of a router and a switch.

8A is a conceptual diagram of a packet that has been randomized according to an embodiment.

As shown in FIG. 8A, the sizes of the payloads 802, 812, 822 and 832 of the first packets X1, X2, X3 and X4 may be randomized and different. Accordingly, the size of each of the first packets X1, X2, X3, and X4 may be different. In addition, although the first packet (X1, X2, X3, X4) in the embodiment of FIG. 8A has been described that the size of the payload is randomized, it is merely an example. Some sizes can be randomized.

8B is a conceptual diagram for explaining a flow processing improvement according to an embodiment of the randomization technique. 8B has an ingress switch environment composed of two input ports (X, Y) and one output port (Z), as in the comparative example of FIG.

As shown in FIG. 8B, the packets to which the randomization technique according to various embodiments are applied may have a random size, respectively. More specifically, since the size of the payload of the packet is set to less than the maximum segment size according to the embodiment, each packet (X1 to X4, Y1 to Y4) may have a random size.

Thus, the processing of the first packet (X1, X2, X3, X4) input to the X input port (Port X) and the second packet (Y1, Y2, Y3, Y4) input to the Y input port , That is, output to the Z output port (Port Z), can be performed fairly in comparison with the comparative example of FIG.

For example, as shown in FIG. 8B, the size of the first packet X2 may be smaller than the size of the second packet Y2. Accordingly, the arrival time T [X2] of the first packet may be earlier than the arrival time T [Y2] of the second packet in contrast to the comparative example of Fig. 2, and the output queue The first packet X2 other than the second packet Y2 may be input. Accordingly, even when the first packets X1 to X4 are synchronized slightly later than the second packets Y1 to Y4, the first packets X1 to X4 and the second packets Y1 to Y4 are appropriately output Can be input to the queue.

9 shows experimental results on a network simulator. First, when the packet size is not randomized (rand = 0), the transmission performance of S1-> R is the lowest. Next, the transmission performance of [S2, S3] -> R was low. However, it can be seen that the transmission performance of all the flows becomes similar when the degree of hydration is increased from 0.8 to 0.9 by gradually increasing the degree of hydration (rand).

On the other hand, the transmitting terminal according to another embodiment can collect a congestion window as randomization related information, and can monitor a congested window, for example. TCP reduces the congestion window when packet loss is detected through the congestion control algorithm. Immediately after decrementing the congested window, the congested window is incrementally increased until the next packet loss is detected, and the increase in the congested window can be performed by an operating system specific algorithm. When the congestion window is reduced immediately after the packet loss is detected, the transmitting terminal can determine that the network congestion is relatively small. Because the flow that detected the loss would have reduced the network congestion because it reduced the sending rate. Conversely, if congestion windows increase, network congestion will increase.

Thus, the transmitting terminal can determine the degree of randomization according to the increase / decrease of the congestion window. For example, FIG. 10A shows an example of a CUBIC, which is a congestion control algorithm used by default in the Linux kernel, and shows the relationship between the time t and the concatenation window cwnd. As shown in Fig. 10A, it can be seen that the congestion window increases with the passage of time.

The transmitting terminal can monitor the congestion window and determine the degree of randomization. For example, FIG. 10B is a graph showing a change in the degree of randomization of a transmitting terminal corresponding to the congestion window state of FIG. 10A. As shown in FIG. 10B, the transmitting terminal may also increase the degree of randomization as the congestion window increases.

It should be understood that the Linux environment 10a described above is merely an example, and those skilled in the art will understand that the randomization degree can be determined using the congestion window regardless of the type of the operating system, Is not limited.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the apparatus and components described in the embodiments may be implemented within a computer system, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA) A programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For ease of understanding, the processing apparatus may be described as being used singly, but those skilled in the art will recognize that the processing apparatus may have a plurality of processing elements and / As shown in FIG. For example, the processing unit may comprise a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as a parallel processor.

The software may include a computer program, code, instructions, or a combination of one or more of the foregoing, and may be configured to configure the processing device to operate as desired or to process it collectively or collectively Device can be commanded. The software and / or data may be in the form of any type of machine, component, physical device, virtual equipment, computer storage media, or device , Or may be permanently or temporarily embodied in a transmitted signal wave. The software may be distributed over a networked computer system and stored or executed in a distributed manner. The software and data may be stored on one or more computer readable recording media.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions to be recorded on the medium may be those specially designed and configured for the embodiments or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI > or equivalents, even if it is replaced or replaced.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (20)

Determining a maximum segment size (MSS) defined in transmitting and receiving a packet to / from a receiving terminal; And
Randomizing the size of the packet to less than the defined maximum segment size
A random number generator for generating a random number; The method according to claim 1,
The step of randomizing the size of the packet comprises:
And determining a degree of randomization of the transmitting terminal. 3. The method of claim 2,
Wherein the step of determining the degree of randomization comprises:
A random number generating method of a transmitting terminal that receives a parameter for adjusting the degree of randomization and determines the degree of randomization based on the received parameter. 3. The method of claim 2,
Wherein the step of determining the degree of randomization comprises:
And the transmitting terminal determines a parameter for adjusting the degree of randomization of the transmitting terminal. 5. The method of claim 4,
Collecting information related to randomization to determine the degree of randomization;
The random number generating method of the transmitting terminal. 6. The method of claim 5,
The randomization related information is at least one of multiplier information and congestion window information. The method according to claim 6,
Wherein the step of determining the degree of randomization comprises:
The randomization technique of the transmitting terminal that collects the multipliers through parameters of the TCP stack of the operating system kernel (TCP stack). The method according to claim 6,
Wherein the step of determining the degree of randomization comprises:
Further comprising the step of determining whether or not to timeout the transmission and reception of the packet in accordance with the variation of the multiplier information. The method according to claim 6,
Wherein the step of determining the degree of randomization comprises:
Further comprising the step of determining whether transmission / reception of the packet is lost corresponding to the variation of the congestion window information. 10. The method according to claim 8 or 9,
Wherein the step of determining the degree of randomization comprises:
If the value of the multiplier is a first value, it is determined that a timeout of transmission / reception of the packet has not occurred, or if a value of the multiplier is a second value, a timeout of transmission / reception of the packet has occurred Randomization technique of transmitting terminal. 11. The method of claim 10,
Wherein the step of determining the degree of randomization comprises:
The random number generating method of the transmitting terminal determines that the timeout of transmission and reception of the packet occurs continuously as the value of the multiplier information increases. 11. The method of claim 10,
Wherein the step of determining the degree of randomization comprises:
The random number generating method of the transmitting terminal determines that loss of transmission / reception of the packet occurs as the value of the congestion window information increases. 11. The method of claim 10,
Wherein the step of determining the degree of randomization comprises:
And the randomization degree is determined according to at least one of the multiplier information and the concatenated window information. 3. The method of claim 2,
Wherein the degree of randomization includes at least one of a maximum value of the randomized value, a minimum value of the randomized value, an expected value of the randomized value, and a standard deviation of the distribution of the randomized value. The method according to claim 1,
The step of randomizing the size of the packet comprises: adjusting a size of a payload of the packet through a randomization function in a transport layer. 16. The method of claim 15,
Transmitting a packet whose size is adjusted in the payload size to a lower layer
The random number generating method of the transmitting terminal. The method according to claim 1,
The step of randomizing the size of the payload of the packet comprises segmenting a segment larger than a maximum segment unit inputted from the upper layer when TSO (Transmit segmentation offload) is activated. The method according to claim 1,
The randomization technique of a transmitting terminal that performs randomization according to a software or hardware defined randomization algorithm in at least one of a router and a switch. Monitoring at least one of multiplier information and congestion window information of a TCP stack of an operating system kernel;
Determining whether the size of a packet to be transmitted to a receiving terminal is randomized based on the monitoring result; And
Transmitting the packet to the receiving terminal
A random number generator for generating a random number; A controller for determining a maximum segment size (MSS) defined when transmitting and receiving a packet to and from a receiving terminal, and for randomizing the packet size to less than the defined maximum segment size; And
And a communication unit for transmitting the packet to the receiving terminal
.

Images