CN101964727B - Method and device for measuring available bandwidth by using mixed messages - Google Patents

Abstract

The invention discloses a method for measuring available bandwidth by using mixed messages. The method includes: the active measurement terminal sends an Internet Control Message Protocol (ICMP) message to the passive measurement terminal, and the active measurement terminal counts the normal ICMP of the passive measurement terminal. Packet loss rate; the active measurement end sends mixed packets to the passive measurement end, and the active measurement end counts the average packet loss rate of ICMP packets in the mixed packets of the passive measurement end; the active measurement end determines the average packet loss rate of ICMP packets When the difference between the packet loss rate of the normal ICMP message and the network is greater than the preset packet loss rate threshold, the average rate of receiving mixed messages at the passive measurement end is counted as the available bandwidth from the active measurement end to the passive measurement end; the invention also discloses a A device for measuring available bandwidth by using mixed packets, based on the method and device can reduce the complexity of bandwidth measurement.

Description

A method and device for measuring available bandwidth using mixed packets

technical field

The invention relates to the field of computer network electronic communication, in particular to a method and device for measuring available bandwidth by using mixed messages.

Background technique

With the development of computer technology, the Internet (Internet) has gradually become one of the important platforms for people to obtain information and communicate with each other. The rapid global expansion of the Internet and the rapid development of new services such as real-time services and multimedia services on the Internet have led to a decline in the performance of the Internet network. At present, the main factors affecting network performance are link bandwidth, link bottleneck, application and protocol design, and host system. Among them, link bandwidth is the most important network resource and one of the main determinants of transmission path performance. , and is closely related to network bottlenecks, overall performance analysis, capacity planning, etc. Therefore, accurate measurement of bandwidth has become an important issue in Internet research.

In the field of network bandwidth measurement, according to different measures, network bandwidth measurement can be divided into different aspects of measurement such as link bandwidth, path bandwidth and available bandwidth. Available bandwidth is a key parameter in network routing, traffic engineering, and QoS control in engineering, and it has more practical significance than bottleneck bandwidth.

Link bandwidth refers to the maximum bandwidth of a link, that is, the maximum data transmission rate of the link, which is determined by the physical properties of the link; link available bandwidth refers to the bandwidth of a link that is not currently being used; path bandwidth is the current path bandwidth The minimum value of the bandwidth of all links on the path, the path bandwidth is also called the end-to-end path bandwidth; the path available bandwidth is the minimum value of the available bandwidth of all links on the current path, and the path available bandwidth is also called the end-to-end path available bandwidth. Among all the links on the current path, the link with the smallest available link bandwidth is called the urgent link of the path, and the link with the smallest link bandwidth is called the bottleneck link of the path.

Figure 1 shows a schematic diagram of the end-to-end bandwidth. The path between the sending end and the receiving end is composed of three links L ₁ , L ₂ and L ₃ , where C ₁ , C ₂ , and C ₃ are link L ₁ , L ₂ , L ₃ link bandwidth; the shaded part indicates the bandwidth already used by the current link, then A ₁ , A ₂ , A ₃ are the unused bandwidth of links L ₁ , L ₂ , L ₃ respectively, namely Link available bandwidth. It can be seen from the figure that C ₁ <C ₂ <C ₃ , that is, the link bandwidth of the link L ₁ is the smallest, which determines the path bandwidth from the sender to the receiver, and the link L ₁ is the link between the sender and the receiver. Bottleneck link; in addition, A ₃ <A ₁ <A ₂ , that is, link L ₃ has the smallest link available bandwidth, which determines the available bandwidth of the path from the sending end to the receiving end, and link L ₃ is an urgent link of the path.

Most of the existing bandwidth measurement methods are packet queuing algorithms based on time stamps, and the main measurement models include the probing interval model (PGM, Packet Gap Model) and probing rate model (PRM, Packet Rate Model). Among them, the algorithm based on the PGM model is mainly to analyze the change of the packet interval of the detection packet queue, and determine the estimated value of the bandwidth according to the change trend, which is also called the bandwidth measurement method based on the time stamp; the algorithm based on the PRM model is mainly at the sending end Send the probe packet queue at a certain rate, and dynamically adjust the sending rate of the probe packet queue according to the rate at which the receiver receives the probe packet queue, so as to determine a matching critical rate to estimate the bandwidth to be tested.

However, the existing bandwidth measurement methods based on PGM and PRM mainly have the following disadvantages:

1. The above two measurement methods generally require the cooperation of the passive measurement end, that is to say, the passive measurement end needs to perform corresponding calculations based on some information of the detection packet queue, and return these calculation information to the active measurement end. The two measurement ends need Only by working together can the link bandwidth be tested. In this case, the passive measurement terminal needs to take auxiliary measures such as installing a corresponding measurement program, which reduces the practicability and flexibility of the measurement method.

2. The calculation process of the above two measurement methods is complicated, especially the PGM-based bandwidth measurement method has relatively strict requirements on timing accuracy and the like, and the calculation complexity is relatively high.

Contents of the invention

In view of this, the main purpose of the present invention is to provide a method and device for measuring available bandwidth by using mixed packets, which can reduce the complexity of bandwidth measurement.

In order to achieve the above object, technical solution of the present invention is achieved in that way:

The present invention provides a method for measuring available bandwidth by using mixed messages, the method comprising:

The active measurement terminal sends an Internet Control Message Protocol ICMP message to the passive measurement terminal, and the active measurement terminal counts the normal ICMP packet loss rate of the network at the passive measurement terminal;

The active measurement end sends mixed messages to the passive measurement end, and the active measurement end counts the ICMP message average packet loss rate of the mixed messages at the passive measurement end;

When the active measurement end judges that the difference between the average packet loss rate of the ICMP message and the normal ICMP packet loss rate of the network is greater than the preset packet loss rate threshold, the average rate at which the passive measurement end receives mixed messages is counted as the active measurement end. The available bandwidth from the measuring end to the passive measuring end;

When the active measurement terminal determines that the difference between the average packet loss rate of ICMP packets and the normal ICMP packet loss rate of the network is less than or equal to a preset packet loss rate threshold, the method further includes: the active measurement terminal re- Determining the sending rate of the mixed message specifically includes: increasing the sending rate in a linear or exponential manner based on the first sending rate of the mixed message.

The method further includes: the active measurement terminal sends an ICMP message to the passive measurement terminal at a preset initial sending rate, and the passive measurement terminal returns the ICMP message to the An active measurement terminal, the active measurement terminal counts the normal ICMP packet loss rate of the network according to the ICMP message sent by itself and the ICMP message returned by the passive measurement terminal.

Wherein, before the active measurement end sends the mixed message to the passive measurement end, the method further includes: the active measurement end determines the sending rate of the mixed message;

Correspondingly, the sending rate of the mixed message is determined, specifically:

When the active measurement end sends the mixed message to the passive measurement end for the first time, the first sending rate of the mixed message is the preset initial sending rate, or the active measurement The link bandwidth of the bottleneck link between peers and passive measurement peers.

The mixed message sent by the active measurement end is n groups of mixed messages, and the mixed message includes the ICMP message and the UDP message, where n is an integer greater than or equal to 1.

The method also includes: the active measurement end sets the destination port of the UDP message to a wrong destination port; correspondingly, after the passive measurement end receives the mixed message, it Discard the UDP message, and return the ICMP message in the mixed message to the active measurement end.

The active measurement terminal counts the average packet loss rate of the ICMP messages, specifically: the active measurement terminal calculates the n groups of mixed messages sent by itself and the ICMP messages returned by the passive measurement terminal. The ICMP packet loss rate of each group of mixed messages is averaged to obtain the average ICMP packet loss rate of the n groups of mixed messages.

The active measurement terminal counts the average rate at which the passive measurement terminal receives the mixed message, specifically: the active measurement terminal calculates the statistics based on the n groups of mixed messages sent by itself and the ICMP messages returned by the passive measurement terminal The passive measurement terminal receives the rate of each group of mixed messages in the n groups, and averages the rates at which the passive measurement terminal receives the n groups of mixed messages to obtain the mixed message received by the passive measurement terminal The average speed of the text.

The present invention also provides a device for measuring available bandwidth by using mixed messages, the device includes: an active measurement terminal and a passive measurement terminal, wherein,

The active measurement terminal is used to send ICMP messages to the passive measurement terminal, and count the normal ICMP packet loss rate of the network at the passive measurement terminal; it is also used to send mixed messages to the passive measurement terminal, And count the average packet loss rate of the ICMP messages at the passive measurement end, and determine that the difference between the average packet loss rate of the ICMP messages and the normal ICMP packet loss rate of the network is greater than the preset packet loss rate When the threshold is reached, count the average rate at which the passive measurement terminal receives the mixed message;

The passive measurement terminal is used to receive the ICMP message and the mixed message;

When the active measurement terminal determines that the difference between the average packet loss rate of ICMP packets and the normal ICMP packet loss rate of the network is less than or equal to the preset packet loss rate threshold: the active measurement terminal re-determines the mixed packet loss rate The sending rate of the message is specifically: based on the first sending rate of the mixed message, the sending rate is increased linearly or exponentially.

Wherein, the passive measurement terminal is also used to return the ICMP message to the active measurement terminal when receiving the ICMP message sent by the active measurement terminal; when receiving the mixed message sent by the active measurement terminal , discard the UDP message in the mixed message, and return the ICMP message in the mixed message to the active measurement end.

In the bandwidth measurement scheme of the present invention, the active measurement terminal sends an ICMP message to the passive measurement terminal, and counts the normal ICMP packet loss rate of the network at the passive measurement terminal; the active measurement terminal sends a mixed message to the passive measurement terminal, and counts The average packet loss rate of ICMP packets to the passive measurement end; when the active measurement end determines that the difference between the average packet loss rate of ICMP packets and the normal ICMP packet loss rate of the network is greater than the preset packet loss rate threshold, the passive measurement end receives statistics. The average rate of mixed packets; the average rate is the measured path available bandwidth. In addition, in this scheme, the destination port of the UDP message in the mixed message sent by the active measurement end is set to be wrong, so that the passive measurement end can directly discard the UDP message and only return the ICMP message to the active measurement end. It can be seen that the bandwidth measurement in the present invention does not require the cooperation of passive measurement terminals, which reduces the complexity of the measurement work, and at the same time, the measurement method of the present invention has relatively high practicability and flexibility.

Description of drawings

Figure 1 is a schematic diagram of end-to-end bandwidth;

FIG. 2 is a schematic flow diagram of a method for measuring available bandwidth using mixed packets in the present invention;

FIG. 3 is a schematic flow chart of a method for measuring available bandwidth using mixed packets according to an embodiment of the present invention;

Fig. 4 is a schematic diagram of the composition of the mixed message of the present invention;

FIG. 5 is a schematic structural diagram of a device for measuring available bandwidth by using mixed packets according to the present invention.

Detailed ways

The technical solutions of the present invention will be further elaborated below in conjunction with the accompanying drawings and specific embodiments.

The bandwidth measurement in the present invention is mainly the measurement of the available bandwidth of the end-to-end path, wherein one end is an active measurement end, and the other end is a passive measurement end; the active measurement end sends a mixed message to the passive measurement end. The available bandwidth of the path at the end is measured.

Mixed message among the present invention refers to User Datagram Protocol (UDP, User Datagram Protocol) message and Internet Control Message Protocol (ICMP, Internet Control Message Protocol) message.

The present invention utilizes the process flow that mixed message measures available bandwidth as shown in Figure 2, comprises:

Step 201, the active measurement end sends an ICMP message to the passive measurement end, and the active measurement end counts the normal ICMP packet loss rate of the network at the passive measurement end;

Step 202, the active measurement end sends mixed messages to the passive measurement end, and the active measurement end counts the average packet loss rate of ICMP packets in the mixed messages of the passive measurement end;

Step 203, when the active measurement end determines that the difference between the average packet loss rate of ICMP packets and the normal ICMP packet loss rate of the network is greater than the preset packet loss rate threshold, the average rate of receiving mixed messages at the passive measurement end is counted as an active measurement The bandwidth available from the end to the passive measurement end.

The method for measuring bandwidth of the present invention is described below in conjunction with specific implementation, as shown in Figure 3, including:

In step 301, the active measurement terminal measures the normal ICMP packet loss rate of the network.

The active measurement end sends ICMP messages to the passive measurement end at the preset initial sending rate. After receiving the ICMP message, the passive measurement end returns the ICMP message to the active measurement end. The ICMP message returned by the passive measurement end, the statistics of the normal ICMP packet loss rate of the network from the active measurement end to the passive measurement end, specifically: the active measurement end sends the ICMP message to the passive measurement end at the preset initial sending rate V ₀ , Among them, V ₀ is a lower transmission rate. When V ₀ is used as the transmission rate of ICMP messages, it is necessary to ensure that no network congestion will be caused. Under this premise, the active measurement terminal can perform V ₀ value according to its own needs. choose.

Assuming that the active measurement end sends z (z is a positive integer) ICMP packets to the passive measurement end at a time, and the passive measurement end returns z'(z' is a positive integer) ICMP packets, then the network on the current path is normal ICMP packet loss rate $\mathrm{\&beta;}=\frac{z-z^{\&prime;}}{z}.$

In step 302, the active measuring end determines the current sending rate of mixed packets.

When the active measurement end sends a mixed message to the passive measurement end for the first time, and the current path, that is, the link bandwidth of the bottleneck link from the active measurement end to the passive measurement end is known, the first sending rate of the mixed message It may be the link bandwidth of the bottleneck link; otherwise, it may be the preset initial sending rate V ₀ .

If the first sending rate of the mixed message cannot meet the measurement of the available bandwidth of the path, the sending rate of the mixed message needs to be increased. Specifically, linear increase and exponential increase can be adopted as required. Among them, when the transmission rate is increased linearly, the growth rate of the transmission rate is slow, resulting in a relatively long test time, but the test results are more accurate; when the transmission rate is increased exponentially, the growth rate of the transmission rate is faster, and the test convergence is faster, but The test results have relatively large errors in increasing the sending rate relatively linearly.

In step 303, the active measurement terminal sends a mixed message to the passive measurement terminal at the currently determined sending rate; the passive measurement terminal returns an ICMP message to the active measurement terminal.

In the mixed message of the present invention, the quantity of UDP message is more, preferably, can set to send k (k is a positive integer) UDP message continuously and then send an ICMP message, as shown in Figure 4 as a group Schematic diagram of the structure of the mixed message. The active sender uses the ICMP ₁ message as the start of the group of mixed messages, and then sends k UDP messages continuously and then sends an ICMP ₂ message; after sending k UDP messages continuously Send an ICMP ₃ message, and so on, send an ICMP _s message after sending k UDP messages continuously as the end of the group of mixed messages, k UDP messages + 1 ICMP message can be called A sending interval for a group of mixed packets. It can be seen that the interval between sending two consecutive ICMP packets is very large, and the interval between passive measurement terminals returning ICMP packets is also very large.

There are s (s is a positive integer) ICMP messages in the mixed message, and the numbers of the ICMP messages are continuous, wherein, ICMP ₁ is the first ICMP message, ICMP _s is the last ICMP message; UDP The number of packets is (s-1)×k.

It should be pointed out that in the present invention, when the active measurement terminal sends the UDP message, the destination port of the UDP message is set to a wrong destination port number, so that when the passive measurement terminal receives the mixed message, due to the UDP message The destination port number is wrong, so the passive measurement end directly discards the UDP message, but only receives the ICMP message, and returns the ICMP message to the active measurement end. Because the ICMP message itself has the function of measuring end-to-end data transmission (including bandwidth measurement), and the UDP message is actively discarded by the passive measuring end, therefore, the passive measuring end does not need to do any cooperative work on the bandwidth measurement, which is greatly improved. It greatly reduces the complexity of bandwidth measurement.

The ratio of the above-mentioned UDP message to the mixed message is far greater than that of the ICMP message to the mixed message. In this way, in the uplink (the active measurement end sends the mixed message to the passive measurement end), due to the data volume of the mixed message Larger, so that the sending of ICMP messages is greatly affected by network conditions; in the downlink (the passive measurement end returns ICMP messages to the active measurement end), since the UDP message has been discarded, only the ICMP message is returned. The amount of data is greatly reduced, and the interval between ICMP messages is large. Therefore, the impact of ICMP messages on network conditions is greatly reduced, which helps to improve the accuracy of bandwidth measurement.

In step 304, the active measurement terminal counts the average packet loss rate of ICMP packets.

The active measurement end counts the ICMP packet loss rate of each group of mixed messages in the n groups according to the n groups of mixed messages sent by itself and the ICMP messages returned by the passive measurement end, and calculates the ICMP packet loss rate of the n groups of mixed messages. Average the packet loss rate of ICMP packets to obtain the average packet loss rate of ICMP packets.

Specifically, the active measurement end needs to count the average packet loss rate of ICMP packets at the current sending rate. When measuring the available bandwidth of a path only based on the ICMP packet loss rate of a single group of mixed messages, a large error will be generated. Therefore, when performing a measurement of the available bandwidth of a path, it is necessary to actively measure the terminal Send n (n=1, 2, 3...) groups of mixed messages at a time, preferably, n can take an integer greater than 1, and calculate the ICMP message packet loss rate of each group of mixed messages, that is, a single Group ICMP packet loss rate α _n ; then take the average of the n groups of ICMP packet loss rates to obtain the average ICMP packet loss rate In this way, the calculation accuracy of the ICMP packet loss rate can be improved.

The calculation of the packet loss rate α _n of a single group of ICMP packets in this step will be described in detail below.

When the active measurement terminal receives the last ICMP message in the group of mixed messages returned by the passive measurement terminal, it confirms whether the received ICMP message is lost; if the ICMP message is lost, the single group in this step The calculation method of the packet loss rate in step 301 can be used for calculating the ICMP packet loss rate α _n .

If the ICMP message is not lost, use the following calculation method:

L为每组混合报文的长度，单位为bit；当某个ICMP报文返回到主动测量端的时间与与第一个ICMP报文返回到主动测量端的时间间隔大于t+Δt时，认为该ICMP报文为丢失的ICMP报文，如此可以得出该组混合报文中丢失的ICMP报文的个数m_n，则

The limit bandwidth (maximum bandwidth) B of this path can be determined according to the current path situation, and the prior art can be adopted here, and will not be repeated; according to the sending time of the ICMP message, the first ICMP message and the last ICMP message can be calculated. The sending interval of the ICMP message is t; according to the time when the ICMP message returns to the active measurement end, the time interval between the first ICMP message and the last ICMP message returning to the active measurement end can be calculated as t+Δt, where,

L is the length of each group of mixed messages, and the unit is bit; when an ICMP message is returned to the active measurement end and the time interval between the first ICMP message and the first ICMP message is returned to the active measurement end is greater than t+Δt, the ICMP The message is a lost ICMP message, so the number m _n of the lost ICMP messages in this group of mixed messages can be obtained, then

Step 305, judging whether the difference between the average ICMP packet loss rate and the normal ICMP packet loss rate of the network is greater than the preset packet loss rate threshold, if not, return to step 302; if yes, execute step 306.

The packet loss rate threshold γ can be set according to the physical characteristics of the current path.

The average packet loss rate of ICMP packets is α, and the packet loss rate of normal ICMP packets is β. Generally, since the sending rate of mixed packets is higher than that of ICMP packets when the network is normal, α must be greater than β, if α-β≤γ, it is determined that the average packet loss rate of ICMP packets at the current sending rate of mixed messages is equivalent to the normal ICMP packet loss rate of the network, indicating that the current sending rate of mixed messages is too low. The current sending rate is based on the fact that the most accurate path available bandwidth cannot be measured. At this time, the active measurement end stops sending mixed messages at the current sending rate, and needs to increase the sending rate of mixed messages. Return to step 302, and the active measuring end restarts. Determine the sending rate of the mixed message, based on the first sending rate of the mixed message, increase the sending rate by linear increase or exponential increase, and then perform step 303, and send the mixed message to the passive measurement end again; It should be pointed out that when the active measurement end sends mixed messages again, it can choose to increase the number of mixed messages sent, such as sending several more sets of mixed messages than last time, and/or increasing the UDP in each set of mixed messages. Number of packets and ICMP packets.

If α-β>γ, it means that the current sending rate of the mixed message does not need to be increased, and it is already a reasonable sending rate for the measurement of the path available bandwidth, and proceeds to step 306 .

In step 306, the active measurement terminal calculates the average rate of mixed packets received by the passive measurement terminal according to the ICMP message returned by the passive measurement terminal, and uses it as the available bandwidth from the active measurement terminal to the passive measurement terminal.

The active measurement terminal counts the rate at which the passive measurement terminal receives each group of mixed messages in the n groups according to the n groups of mixed messages sent by itself and the ICMP messages returned by the passive measurement terminal, and calculates the rate at which the passive measurement terminal receives the n groups of mixed messages. The rate of the packets is averaged to obtain the average rate of mixed packets received by the passive measurement end.

则该平均速率即为测量得到的当前路径、即主动测量端到被动测量端的可用带宽。Specifically, the active measurement terminal needs to measure the average rate at which the passive measurement terminal receives n (n=1, 2, 3...) groups of mixed messages at the current sending rate. Specifically, the active measurement terminal first calculates the passive measurement The rate V _n of receiving a single group of mixed messages at the terminal, and then count the average rate V of these n groups of rates, that is, the average rate at which the passive measurement terminal receives mixed messages

Then the average rate is the measured current path, that is, the available bandwidth from the active measurement end to the passive measurement end.

The statistics of the rate at which the passive measurement terminal receives a single group of mixed packets from the active measurement terminal is based on the premise that the ICMP packets returned by the passive measurement terminal are not affected by the network environment. However, in actual applications, the ICMP packets are basically affected by the network environment. Therefore, the active measurement terminal needs to sample the returned ICMP packets, and select the ICMP packets that are least affected by the network environment as a basis to count the single group of mixed packets received by the passive measurement terminal. text rate.

Concrete ICMP message sample filtering of the present invention adopts following method:

1. Assume that the active measurement terminal receives the ICMP message returned by the passive measurement terminal with the minimum number i and the maximum number j, and the time for the minimum number ICMP message and the maximum number ICMP message to return to the active measurement terminal are respectively T _i and T _j , then the average time interval for the ICMP message to return to the active measurement terminal is $\mathrm{\&Delta;T}=\frac{T_j-T_i}{j-i};$

2. The ideal time for the ICMP message numbered x (i<x≤j) to return to the active measurement terminal is T _x ′=T _i +(xi)×ΔT;

3. The active measurement terminal selects the ICMP message that is closest to T _x (the time when the ICMP message that is numbered x returns to the active measurement terminal) and T _x ', that is, the ICMP message that the time returned to the active measurement terminal is closest to the ideal time, It is used as a basis to calculate the rate at which the passive measurement terminal receives the group of mixed messages.

Of course, other sample filtering algorithms in the prior art may also be used for sample filtering of ICMP messages.

Assuming that the length of the ICMP message in this group of mixed messages is _lICMP , the length of the UDP message is _lUDP , and the unit is bit, then the rate at which the passive measurement terminal receives this group of mixed messages is: $V_{\mathrm{no}}=\frac{(x-i)\&times;k\&times;l_{\mathrm{UDP}}+(x-i+1)\&times;l_{\mathrm{ICMP}}}{T_x-T_i},$ Where k is the number of UDP packets in one sending interval of a group of mixed packets.

In order to realize the above-mentioned bandwidth measurement method, the present invention also provides a measurement device, as shown in FIG. 5 , including: an active measurement terminal 10 and a passive measurement terminal 20, wherein,

The active measurement terminal 10 is used to send ICMP messages to the passive measurement terminal 20, and counts the normal ICMP packet loss rate of the network at the passive measurement terminal 20; it is also used to send mixed messages to the passive measurement terminal 20, and counts to The average packet loss rate of ICMP packets at the passive measurement terminal 20; it is also used to count the passive measurement terminal when the difference between the average packet loss rate of ICMP packets and the normal ICMP packet loss rate of the network is greater than the preset packet loss rate threshold 20 The average rate of receiving mixed messages;

The passive measurement terminal 20 is used for receiving ICMP messages and mixed messages. It is also used to return the ICMP message to the active measurement terminal 10 when receiving the ICMP message sent by the active measurement terminal 10; when receiving the mixed message sent by the active measurement terminal 10, the UDP in the mixed message The message is discarded, and the ICMP message in the mixed message is returned to the active measurement terminal 10 .

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention.

Claims (9)

1. A method utilizing mixed packets to measure available bandwidth, characterized in that the method comprises: The active measurement terminal sends an Internet Control Message Protocol ICMP message to the passive measurement terminal, and the active measurement terminal counts the normal ICMP packet loss rate of the network at the passive measurement terminal; The active measurement end sends mixed messages to the passive measurement end, and the active measurement end counts the ICMP message average packet loss rate of the mixed messages at the passive measurement end; When the active measurement end judges that the difference between the average packet loss rate of the ICMP message and the normal ICMP packet loss rate of the network is greater than the preset packet loss rate threshold, the average rate at which the passive measurement end receives mixed messages is counted as the active measurement end. The available bandwidth from the measuring end to the passive measuring end; When the active measurement terminal determines that the difference between the average packet loss rate of ICMP packets and the normal ICMP packet loss rate of the network is less than or equal to a preset packet loss rate threshold, the method further includes: the active measurement terminal re- Determining the sending rate of the mixed message specifically includes: increasing the sending rate in a linear or exponential manner based on the first sending rate of the mixed message. 2. The method for measuring available bandwidth using mixed messages according to claim 1, wherein the method further comprises: the active measurement terminal sends an ICMP message to the passive measurement terminal at a preset initial sending rate, and the After the passive measurement terminal receives the ICMP message, it returns the ICMP message to the active measurement terminal, and the active measurement terminal calculates the ICMP message according to the ICMP message sent by itself and the ICMP message returned by the passive measurement terminal. Describes the normal ICMP packet loss rate on the network. 3. The method for measuring available bandwidth using mixed messages according to claim 2, characterized in that, before the active measurement end sends the mixed message to the passive measurement end, the method further comprises: the active measurement end determines The sending rate of the mixed message; Correspondingly, the sending rate of the mixed message is determined, specifically: When the active measurement end sends the mixed message to the passive measurement end for the first time, the first sending rate of the mixed message is the preset initial sending rate, or the active measurement The link bandwidth of the bottleneck link between peers and passive measurement peers. 4. according to claim 1, utilize mixed message to measure the method for available bandwidth, it is characterized in that, the mixed message sent by the active measuring end is n groups of mixed messages, and the mixed message includes the ICMP message text and User Datagram Protocol UDP packets, where n is an integer greater than or equal to 1. 5. according to claim 4, utilize mixed message to measure the method for available bandwidth, it is characterized in that, the method also comprises: described active measuring terminal sets the purpose port of described UDP message to wrong purpose port; Corresponding After receiving the mixed message, the passive measurement end discards the UDP message according to the wrong destination port, and returns the ICMP message in the mixed message to the active measurement end. 6. according to claim 5, utilize mixed message to measure the method for usable bandwidth, it is characterized in that, described active measurement terminal counts described ICMP message average packet loss rate, specifically: described active measurement terminal sends according to self n groups of mixed messages and the ICMP messages returned by the passive measurement end, count the ICMP packet loss rate of each group of mixed messages in the n groups, and discard the ICMP messages of the n groups of mixed messages Rates are averaged to obtain the average packet loss rate of the ICMP packets. 7. The method for measuring available bandwidth using mixed messages according to claims 1 to 6, wherein the active measurement terminal counts the average rate at which the passive measurement terminal receives the mixed messages, specifically: the The active measurement end counts the rate at which the passive measurement end receives each group of mixed messages in the n groups according to the n groups of mixed messages sent by itself and the ICMP messages returned by the passive measurement end, and calculates the rate at which the passive measurement end receives each group of mixed messages in the n groups averaging the rates at which the n groups of mixed messages are received, to obtain an average rate at which the passive measurement end receives the mixed messages. 8. A device for measuring available bandwidth using mixed messages, characterized in that the device comprises: an active measurement terminal and a passive measurement terminal, wherein, The active measurement terminal is used to send ICMP messages to the passive measurement terminal, and count the normal ICMP packet loss rate of the network at the passive measurement terminal; it is also used to send mixed messages to the passive measurement terminal, And count the average packet loss rate of the ICMP messages at the passive measurement end, and determine that the difference between the average packet loss rate of the ICMP messages and the normal ICMP packet loss rate of the network is greater than the preset packet loss rate When the threshold is reached, count the average rate at which the passive measurement terminal receives the mixed message; The passive measurement terminal is used to receive the ICMP message and the mixed message; When the active measurement terminal determines that the difference between the average packet loss rate of ICMP packets and the normal ICMP packet loss rate of the network is less than or equal to the preset packet loss rate threshold: the active measurement terminal re-determines the mixed packet loss rate The sending rate of the message is specifically: based on the first sending rate of the mixed message, the sending rate is increased linearly or exponentially. 9. The device for measuring available bandwidth using mixed messages according to claim 8, wherein the passive measurement terminal is also used for returning the ICMP message when receiving the ICMP message sent by the active measurement terminal. message to the active measurement end; when receiving the mixed message sent by the active measurement end, discard the UDP message in the mixed message, and return the ICMP message in the mixed message to the active measurement end .

Images