CN104660525A - Bandwidth allocation method, controller and communication system - Google Patents

Abstract

The present invention relates to the field of communication technology, and provides a bandwidth allocation method, comprising: a controller receives data stream bandwidth request information and data stream priority information sent by each host; and according to the received data stream bandwidth request information and data stream The priority information, as well as the ingress port bandwidth information and the egress port bandwidth information of each host, generate bandwidth allocation information; and send the bandwidth allocation information to each host. Correspondingly, the embodiment of the present invention also provides a controller and a communication system, so that high-priority data streams can be responded in a timely manner, and the bandwidth resources of each host can be allocated reasonably, without causing congestion due to over-allocation, Insufficient utilization of bandwidth resources will not be caused by too little allocation, which avoids waste of bandwidth resources.

Description

A bandwidth allocation method, controller and communication system

【Technical field】

The invention relates to the technical field of communication, in particular to a bandwidth allocation method, a controller and a communication system.

【Background technique】

In the data center, the TCP protocol is generally used for data transmission. The TCP protocol has its own congestion control mechanism. When it finds that the data transmitted by a stream is congested, it will slow down the sending rate by adjusting the size of the sending window and the congestion window of the stream, thereby effectively avoiding congestion due to congestion. packet loss and improve the efficiency of data transmission.

However, if the transmission rate of data is not limited and only the congestion control mechanism of TCP is used to coordinate the transmission of flows in the data center, TCP will eventually allocate the bandwidth fairly to the flows in the data center. However, streams in a data center often have different priorities, and streams with higher priority should be transmitted first, such as streams from user queries than background traffic (such as streams generated by internal staff for scientific computing in a company). Get a more timely response. In order to transmit the streams in the data center with different priorities to meet different needs, before transmitting the data, the data can be transmitted to an agent first, and the agent will process it according to certain rules and then use TCP for transmission.

As shown in Figure 1, host M ₃ sends data to three hosts M ₆ , M ₇ , and M ₈ at the same time, and host M ₇ also sends data to five hosts M ₁ , M ₂ , M ₃ , M ₄ , and M ₅ data. In the prior art, each host will evenly distribute its egress bandwidth to the flows originating from this host. For example, the egress bandwidth of the host M ₃ is r ₃ , then the flows from M ₃ to M ₆ , M ₇ , and M ₈ can all be transmitted at a rate of up to r ₃ /3. Similarly, if the egress bandwidth of host M ₇ is r ₇ , each flow from M ₇ can transmit data at a rate of r ₇ /5 at most.

All streams get the same bandwidth as stream priority is not taken into account. In this case, some streams with higher priority may not get a timely response. For example, when the company's researchers are conducting scientific experiments and the data center is transmitting a large amount of data, requests from ordinary users come. However, because all streams have the same priority in technology, the streams of ordinary users do not get a timely response, which will seriously affect the user experience in actual products and bring losses to the company. In addition, unilaterally assigning egress bandwidth to a flow may cause congestion at the destination host or fail to fully utilize the ingress bandwidth of the destination host, resulting in a waste of bandwidth resources. For example, in Figure 1, it is assumed that the egress bandwidth of M ₃ and M The ingress bandwidth of ₇ is r, then M ₃ will send data to M ₇ with a bandwidth of r/3 at most, and M ₇ can only process the flow from M ₃ at a rate of r/5 at most, so packet loss occurs phenomenon, resulting in a waste of bandwidth resources.

【Content of invention】

The invention provides a method, a controller and a communication system for bandwidth allocation based on global information such as data flow priority, egress bandwidth of a source host, and ingress bandwidth of a destination host.

An embodiment of the present invention provides a bandwidth allocation method, including:

The controller receives the data flow bandwidth request information and data flow priority information sent by each host;

generating bandwidth allocation information according to the received data flow bandwidth request information and data flow priority information, and the incoming port bandwidth information and outgoing port bandwidth information of each host;

sending the bandwidth allocation information to each host.

The embodiment of the present invention also provides a controller, including:

A receiving module, configured to receive data flow bandwidth request information and data flow priority information sent by each host;

A bandwidth allocation module, configured to generate bandwidth allocation information according to the data stream bandwidth request information and data stream priority information received by the receiving module, and the ingress port bandwidth information and egress port bandwidth information of each host;

A sending module, configured to send the bandwidth allocation information to each of the hosts.

In addition, an embodiment of the present invention also provides a communication system, including a controller and at least two hosts, the controller is configured to receive data flow bandwidth request information and data flow priority information sent by each host;

generating bandwidth allocation information according to the received data flow bandwidth request information and data flow priority information, and the incoming port bandwidth information and outgoing port bandwidth information of each host;

sending the bandwidth allocation information to each host.

Compared with the prior art, the present invention has the beneficial effects that: the embodiment of the present invention performs bandwidth allocation according to the data flow bandwidth request information and data flow priority information, as well as the ingress port bandwidth information and the egress port bandwidth information, which can The high-priority data flow can be responded in a timely manner, and the bandwidth resources of each host can be allocated reasonably, neither causing congestion due to excessive allocation, nor causing insufficient bandwidth resource utilization due to too little allocation, avoiding Waste of bandwidth resources.

【Description of drawings】

FIG. 1 is a topological diagram of a data stream transmission network in the prior art;

Fig. 2 is the method flowchart of embodiment 1 of the present invention;

FIG. 3 is a topological diagram of a data stream transmission network according to Embodiment 2 of the present invention;

FIG. 4 is an information diagram of data stream transmission in Embodiment 3 of the present invention;

Fig. 5 is a structural block diagram of a device according to Embodiment 4 of the present invention.

【Detailed ways】

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.

In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not constitute a conflict with each other.

Embodiment one:

Embodiment 1 of the present invention provides a bandwidth allocation method, as shown in FIG. 2 , including:

S101. The controller receives data flow bandwidth request information and data flow priority information sent by each host;

Here, the number of hosts is m. Each host sends the data flow bandwidth request sent by the host as the source host to the destination host and the priority of the data flow to the controller.

S102. Generate bandwidth allocation information according to the received data stream bandwidth request information and data stream priority information, and the ingress port bandwidth information and egress port bandwidth information of each host;

Since the bandwidth of the ingress port and the bandwidth of the egress port of each host are constant, the controller knows the bandwidth of the ingress port and the bandwidth of the egress port of each host in advance, and the bandwidth information of the ingress port of the host M _i is expressed as R(P _i ⁱⁿ ) , and the outbound bandwidth information is denoted as R(P _i ^out ).

After the controller receives the data flow bandwidth request information and data flow priority information sent by each host, it can generate a data flow bandwidth request matrix R ⁰ and a priority weight matrix W, wherein the i-th row and the jth column of the R ⁰ element is the bandwidth request of the data flow f _ij of the host M _i to M _j , and the i-th row j column element ω _ij of the priority weight matrix W is the priority weight of the data flow f _ij of the host M _i to M _j value, and the R ⁰ and W are both m×m matrices, m represents the number of hosts, and i and j are natural numbers.

Specifically, for each row of the matrix R ⁰ , first calculate the bandwidth allocated by the data flow f _ij according to the priority weight in $r_{\mathrm{ij}}^{\mathrm{temp}}=R\left({P_i}^{\mathrm{out}}\right)\×\frac{{\mathrm{\ω}}_{\mathrm{ij}}}{\underset{j=1,...,m}{\mathrm{\Σ}}{\mathrm{\ω}}_{\mathrm{ij}}};$

then and The smaller value of the two is taken as the element ^of row i, column j in matrix R1

According to the remaining bandwidth of the outgoing port bandwidth R(P _i ^out ) of the host M _i , update the elements in the i-th row and j-th column in the matrix R ¹

For each column of the matrix R ¹ , the bandwidth allocated by the data flow f _ij according to the priority weight is calculated in turn in $r_{\mathrm{ij}}^{\mathrm{temp}}=R\left({P_i}^{\mathrm{in}}\right)\×\frac{{\mathrm{\ω}}_{\mathrm{ij}}}{\underset{i=1,...,m}{\mathrm{\Σ}}{\mathrm{\ω}}_{\mathrm{ij}}};$

Will and The smaller of the ^two is taken as the element at row i, column j in matrix R2

_{Update _} ^_ After each element is updated, bandwidth allocation information can be obtained.

S103. Send the bandwidth allocation information to each host.

Specifically, the bandwidth allocation information related to the host can be sent to each host or the matrix R ² information can be sent to each host, wherein the i-th row and j-th column element is the bandwidth value of data flow f _ij from host M _i to M _j .

In the embodiment of the present invention, bandwidth allocation is performed according to the data stream bandwidth request information and data stream priority information, as well as the ingress port bandwidth information and egress port bandwidth information, so that high priority data streams can be responded in a timely manner, and each host The bandwidth resources can be allocated reasonably, neither causing congestion due to over-allocation, nor underutilization of bandwidth resources due to too little allocation, avoiding the waste of bandwidth resources.

Embodiment two:

Embodiment 2 of the present invention provides a bandwidth allocation method, and its data stream transmission topology diagram is shown in FIG. 3 , including one controller and three hosts. The difference from Embodiment 1 is that this embodiment is aimed at the situation where the number of hosts is three, and its bandwidth allocation method is similar to Embodiment 1, specifically:

The controller receives the data flow bandwidth request information and data flow priority information sent by the three hosts;

In a data center composed of three hosts, there are six data streams to be transmitted, which are recorded as f ₁₂ , f ₁₃ , f ₂₁ , f ₂₃ , f ₃₁ and f ₃₂ , and f _ij means from host M _i to M _j Each data flow has its bandwidth request and priority. Assign a weight to each data stream according to its priority, and the higher the priority of a data stream, the higher its weight. ω _ij is the priority weight of data flow f _ij from host M _i to M _j , is the bandwidth request of data flow f _ij from host M _i to M _j , Indicates that there is no data flow between host M _i and M _j . What the embodiments of the present invention do is to allocate bandwidth to each data flow according to its priority under the condition that the total bandwidth is limited.

Since the ingress bandwidth and egress bandwidth of each host are constant, the controller knows the ingress bandwidth and egress bandwidth of each host in advance. After the controller receives the data flow bandwidth request value and the priority weight value sent by each host, it establishes a bandwidth matrix and a priority weight value matrix, which are respectively recorded as R ⁰ and W:

${\mathrm{<span\; class="notranslate">\; R</span>}}^{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}^{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 13</span>}}^{\mathrm{<span\; class="notranslate">\; 0</span>}}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}^{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; three</span>}}^{\mathrm{<span\; class="notranslate">\; 0</span>}}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}^{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 32</span>}}^{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right)$

$\mathrm{<span\; class="notranslate">\; W</span>}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 11</span>}}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 13</span>}}\\ {\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; two</span>}}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; three</span>}}\\ {\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 32</span>}}{\mathrm{<span\; class="notranslate">\; \&omega;</span>}}_{\mathrm{<span\; class="notranslate">\; 33</span>}}\end{array}\right)$

Generate bandwidth allocation information according to the received data flow bandwidth request information and data flow priority information, and the incoming port bandwidth information and outgoing port bandwidth information of the three hosts;

Specifically, first update the values of the elements of the matrix R ⁰ row by row, that is, the bandwidth that each flow should be allocated to. The first row of the matrix R ⁰ has two flows f ₁₂ and f ₁₃ , and their weights are ω ₁₂ and ω ₁₃ respectively. The egress bandwidth of host M ₁ is R(P ₁ ^out ), so it is allocated according to the proportion of priority weight, and the bandwidth that should be allocated to flow f ₁₂ is $r_{12}^{\mathrm{temp}}=R\left({P_1}^{\mathrm{out}}\right)\&Center\; Dot;\frac{{\mathrm{\&omega;}}_{12}}{{\mathrm{\&omega;}}_{12}+{\mathrm{\&omega;}}_{13}},$ The bandwidth that should be allocated to flow f ₁₃ is $r_{13}^{\mathrm{temp}}=R\left({P_1}^{\mathrm{out}}\right)\&Center\; Dot;\frac{{\mathrm{\&omega;}}_{13}}{{\mathrm{\&omega;}}_{12}+{\mathrm{\&omega;}}_{13}}.$ Then compare the bandwidth that should be allocated to each flow with the bandwidth requested by this flow.

if Indicates that the allocated bandwidth of flow f ₁₂ cannot meet or just meet its bandwidth request, at this time, all the allocated bandwidth is allocated to it, so that immediately and The smaller value of the two is taken as the element ^of row i, column j in matrix R1 Conversely, if Indicates that the bandwidth allocated to flow f ₁₂ is greater than its bandwidth request, so there is no need to allocate additional bandwidth to it in addition to its bandwidth request, so keep unchanged, ie

Use the same method to update the bandwidth of flow f _13. After processing the data flow corresponding to this line, check whether the output port of M ₁ has additional remaining bandwidth. According to the remaining bandwidth of the outgoing port bandwidth R(P ₁ ^out ) of M ₁ , update the element in row i and column j in matrix R ¹ Specifically: for example That is, it is calculated that the bandwidth that should be allocated to flow f ₁₂ is not fully allocated to it, so there is a certain amount of bandwidth remaining, and the remaining bandwidth is At this time, check whether the flow of this row still needs bandwidth. Assume that the allocated bandwidth of f ₁₃ is less than its bandwidth request, if Allocate the remaining bandwidth Rem(P ₁ ^out ) to the flow f ₁₃ , otherwise only allocate the part of the remaining bandwidth Rem(P ₁ ^out ) to the flow f ₁₃ to satisfy its bandwidth request, such that Perform the same operation on each row of the matrix R ⁰ , thus completing the bandwidth allocation of the outgoing port and obtaining a new bandwidth matrix

${\mathrm{<span\; class="notranslate">\; R</span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 13</span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; three</span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 32</span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right)$

These streams are then allocated bandwidth on a column-by ^- column basis on the basis of R1. Since each column corresponds to an ingress port of a host, allocating bandwidth by column is equivalent to allocating bandwidth to ingress ports of a host. Take the first column of R ¹ as an example, this column contains two flows f ₂₁ and f ₃₁ . In the process of bandwidth allocation to the egress ports, the bandwidth is allocated to them respectively and It is known that the ingress port bandwidth of M ₁ is R(P ₁ ⁱⁿ ), and the priority weights of flows f ₂₁ and f ₃₁ are ω ₂₁ and ω ₃₁ respectively, so according to the ratio of priority weights, the bandwidths that should be allocated to them are respectively for $r_{\mathrm{twenty\; one}}^{\mathrm{temp}}=R\left({P_1}^{\mathrm{in}}\right)\&CenterDot;\frac{{\mathrm{\&omega;}}_{\mathrm{twenty\; one}}}{{\mathrm{\&omega;}}_{\mathrm{twenty\; one}}+{\mathrm{\&omega;}}_{31}}$ and $r_{31}^{\mathrm{temp}}=R\left({P_1}^{\mathrm{in}}\right)\&Center\; Dot;\frac{{\mathrm{\&omega;}}_{31}}{{\mathrm{\&omega;}}_{\mathrm{twenty\; one}}+{\mathrm{\&omega;}}_{31}}.$

Note that when allocating bandwidth by column, the bandwidth allocated to each flow cannot exceed the bandwidth allocated by them when allocating bandwidth by row, because the bandwidth allocated to a flow when allocating bandwidth by row is the bandwidth of this flow The maximum bandwidth that the source host can support. Consider first updating the bandwidth allocated to flow f ₂₁ , if Indicates that the bandwidth that the destination host of flow f ₂₁ can support is less than or equal to the bandwidth that its source host can allocate to it, then the bandwidth allocated to flow f ₂₁ is updated as Conversely, if Indicates that the bandwidth allocated by the destination host of flow f ₂₁ is greater than the bandwidth allocated by it when it is allocated by row, and the bandwidth allocated by each flow when it is allocated by column cannot exceed the bandwidth allocated by them when it is allocated by row , so keeping its bandwidth constant, ie

After allocating bandwidth to the two flows in the first column, if the sum of their bandwidths is less than the bandwidth of the destination port P ₁ ⁱⁿ , for example, because That is, the bandwidth originally allocated to flow f ₂₁ is not fully allocated, thus $r_{\mathrm{twenty\; one}}^2+r_{31}^2<R\left({P_1}^{\mathrm{in}}\right),$ At this time, the remaining bandwidth is obtained $\mathrm{Rem}\left({P_1}^{\mathrm{in}}\right)=R\left({P_1}^{\mathrm{in}}\right)-r_{31}^2-r_{31}^2.$ If it is judged that the bandwidth allocated to flow f ₃₁ at this stage is less than the bandwidth allocated when it is allocated by row, that is Since there is remaining bandwidth, the remaining bandwidth can be allocated to flow f ₃₁ . if Then all the remaining bandwidth of the ingress port of M ₁ is allocated to flow f ₃₁ , namely Otherwise, allocate part of the remaining bandwidth of the ingress port of M ₁ to flow f ₃₁ such that Treat the second and third columns of R ¹ in the same way to get the bandwidth matrix

${\mathrm{<span\; class="notranslate">\; R</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 13</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; one</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; twenty\; three</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 31</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 32</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right)$

Send the bandwidth allocation information to the above three hosts.

Among them, the controller can send matrix R ² information to each host, in which the i-th row and j-th column element is the bandwidth value of the data flow f _ij of the host M _i to M _j ; the bandwidth information related to the host can also be sent to each host, that is, only some elements in the bandwidth allocation matrix are sent to the host. After receiving the bandwidth allocation information, the host performs final bandwidth allocation according to the information.

In the embodiment of the present invention, bandwidth allocation is performed according to the data stream bandwidth request information and data stream priority information, as well as the ingress port bandwidth information and egress port bandwidth information, so that high priority data streams can be responded in a timely manner, and each host The bandwidth resources can be allocated reasonably, neither causing congestion due to over-allocation, nor underutilization of bandwidth resources due to too little allocation, avoiding the waste of bandwidth resources.

Embodiment three:

Embodiment 3 of the present invention provides a specific example of bandwidth allocation for three hosts. FIG. 4 is a data stream transmission information diagram of Embodiment 3 of the present invention. The bandwidth allocation method is similar to Embodiment 1, specifically:

The controller receives the data flow bandwidth request information and data flow priority information sent by the three hosts. bandwidth request here Their priority weights are respectively ω ₁₂ =1, ω ₁₃ =4, ω ₂₁ =2, ω ₂₃ =3, ω ₃₁ =5, ω ₃₂ =7, $R\left({P_1}^{\mathrm{out}}\right)=6,R\left({P_2}^{\mathrm{out}}\right)=8,R\left({P_3}^{\mathrm{out}}\right)=4,R\left({P_1}^{\mathrm{in}}\right)=7,$ $R\left(P_2^{\mathrm{in}}\right)=9,R\left(P_3^{\mathrm{in}}\right)=3.$ There are 6 streams in the data center, and the bandwidth request matrix and priority weight matrix of these streams are respectively

${\mathrm{<span\; class="notranslate">\; R</span>}}^{\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; 0,2,3</span>}\\ \mathrm{<span\; class="notranslate">\; 2,0,4</span>}\\ \mathrm{<span\; class="notranslate">\; 1,5,0</span>}\end{array}\right)$

$\mathrm{<span\; class="notranslate">\; W</span>}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; 0</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 1,4</span>}\\ \mathrm{<span\; class="notranslate">\; 2,0</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 3</span>}\\ \mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right)$

First, bandwidth is allocated to these streams on a row-by-row basis. easy to calculate and So the bandwidth of flow f ₁₂ can be updated as can also be calculated and So the bandwidth of flow f ₁₃ can be kept constant, namely Subsequently, check the remaining bandwidth _of the egress port of M1 And the bandwidth 1.2 allocated to flow f ₁₂ is smaller than its bandwidth request 2, so 0.8 can be allocated to it from the remaining bandwidth. Then for the flow in the outbound port of M ₁ , the final allocation situation is Using the same method, process the second and third rows of R ⁰ to obtain the bandwidth matrix

${\mathrm{<span\; class="notranslate">\; R</span>}}^{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; 0,2,3</span>}\\ \mathrm{<span\; class="notranslate">\; 2,0,4</span>}\\ \mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; 0</span>}\end{array}\right)$

Then, update ^R1 by column. For example, for the first column, because So keep the bandwidth allocated to flow f ₂₁ unchanged. because That is, the bandwidth allocated to flow f ₃₁ when allocated by column is greater than the bandwidth allocated to it when allocated by row, but since the bandwidth allocated to a flow by column should not be greater than the bandwidth allocated to this flow by row, it remains allocated to The bandwidth of flow f ₃₁ is constant, i.e. Then carry out bandwidth allocation for the ingress port of _M2 , which is easy to calculate, and So the ingress port _of host _M2 can only allocate bandwidth 1.1 to flow f12, so same, and However, since the bandwidth allocated to a flow by column should not be greater than the bandwidth allocated to this flow by row, the bandwidth of f ₃₂ remains unchanged, ie

Then check the remaining bandwidth, because the sum of the bandwidth allocated by f ₁₂ and f ₃₂ is less than the bandwidth of the ingress port of M ₂ , the remaining bandwidth of the ingress port of M ₂ is The bandwidth allocated by flow f ₁₂ when allocated by row is But now the allocated bandwidth is only Therefore 0.9 can be allocated to flow f ₁₂ from the remaining bandwidth, so Use the same method to allocate bandwidth to the third column of R ¹ , and finally get the bandwidth allocation matrix

${\mathrm{<span\; class="notranslate">\; R</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\left(\begin{array}{c}\mathrm{<span\; class="notranslate">\; 0,2,1.7</span>}\\ \mathrm{<span\; class="notranslate">\; 2,0,1.3</span>}\\ \mathrm{<span\; class="notranslate">\; 1,3,0</span>}\end{array}\right)$

The controller sends matrix R ² information to each host, in which the i-th row and j-th column element is the bandwidth value of data flow f _ij from host M _i to M _j . It is also possible to send bandwidth information related to the host to each host, that is, only send some elements in the bandwidth allocation matrix to the host. After receiving the bandwidth allocation information, the host performs final bandwidth allocation according to the information.

In the embodiment of the present invention, bandwidth allocation is performed according to the data stream bandwidth request information and data stream priority information, as well as the ingress port bandwidth information and egress port bandwidth information, so that high priority data streams can be responded in a timely manner, and each host The bandwidth resources can be allocated reasonably, neither causing congestion due to over-allocation, nor underutilization of bandwidth resources due to too little allocation, avoiding the waste of bandwidth resources.

Embodiment four:

Embodiment 4 of the present invention provides a controller. FIG. 5 is a structural block diagram of a device according to Embodiment 4 of the present invention, including a receiving module 501, a bandwidth allocation module 502, and a sending module 503, wherein

The receiving module 501 is configured to receive data flow bandwidth request information and data flow priority information sent by each host;

The bandwidth allocation module 502 is configured to generate bandwidth allocation information according to the data flow bandwidth request information and data flow priority information received by the receiving module 501, as well as the ingress port bandwidth information and the egress port bandwidth information of each host;

The sending module 503 is configured to send the bandwidth allocation information to the hosts.

Specifically, the controller may also include a generating module, configured to generate a data stream bandwidth request matrix R ⁰ and a priority weight matrix W, wherein the i-th row and j-th column element of the R ⁰ is the bandwidth request of the data flow f _ij of the host M _i to M _j , and the i-th row j column element ω _ij of the priority weight matrix W is the priority weight of the data flow f _ij of the host M _i to M _j value, the R ⁰ and W are both m×m matrices, m represents the number of hosts, i and j are natural numbers, and the i≤m, j≤m.

Wherein, the bandwidth allocation module 502 is specifically used for:

For each row of the matrix R ⁰ , sequentially calculate the bandwidth allocated by the data flow f _ij according to the priority weight in R(P _i ^out ) is the bandwidth information of the outgoing port of the host M _i ;

Will and The smaller value of the two is taken as the element ^of row i, column j in matrix R1

According to the remaining bandwidth of the outgoing port bandwidth R(P _i ^out ) of the host M _i , update the elements in the i-th row and j-th column in the matrix R ¹

For each column of the matrix R ¹ , the bandwidth allocated by the data flow f _ij according to the priority weight is calculated in turn in R(P _i ⁱⁿ ) is the ingress port bandwidth information of the host M _i ;

Will and The smaller of the ^two is taken as the element at row i, column j in matrix R2

_{Update _} ^_

For the method for obtaining the bandwidth matrix, reference may be made to Embodiment 2 and Embodiment 3, and details are not repeated here.

The sending module 503 is specifically configured to send bandwidth allocation information related to the host to each host; or send matrix R ² information to each host, wherein the i-th row and j-th column element is the bandwidth value of data flow f _ij from host M _i to M _j .

The controller provided by the embodiment of the present invention performs bandwidth allocation according to the data stream bandwidth request information and data stream priority information, as well as the ingress port bandwidth information and the egress port bandwidth information, so that high priority data streams can be obtained in a timely manner. In response, the bandwidth resources of each host can be allocated reasonably, neither causing congestion due to over-allocation, nor underutilization of bandwidth resources due to too little allocation, avoiding the waste of bandwidth resources.

Embodiment five:

Embodiment 5 of the present invention provides a communication system, and its network architecture diagram is similar to FIG. 3 , except that there may be more than three hosts in this embodiment.

The communication system provided by the embodiment of the present invention includes a controller and at least two hosts, the controller is used for:

Receive data flow bandwidth request information and data flow priority information sent by each host;

generating bandwidth allocation information according to the received data flow bandwidth request information and data flow priority information, and the incoming port bandwidth information and outgoing port bandwidth information of each host;

Send the bandwidth allocation information to each host.

It should be pointed out that the controller here is the same as that of the fourth embodiment, and its implementation method is also the same as that of the above method embodiment. The controller of the communication system may be a switch or a certain host, which is not limited in the present invention.

In the communication system provided by the embodiment of the present invention, bandwidth allocation is performed according to the data stream bandwidth request information and data stream priority information, as well as the ingress port bandwidth information and egress port bandwidth information, so that high priority data streams can be obtained in a timely manner. In response, the bandwidth resources of each host can be allocated reasonably, neither causing congestion due to over-allocation, nor underutilization of bandwidth resources due to too little allocation, avoiding the waste of bandwidth resources.

Those of ordinary skill in the art can understand that all or part of the process of the method in the above embodiments can be completed by hardware related to computer program instructions. The program can be stored in a computer-readable storage medium. When the program is executed, it can include Such as the flow of the embodiments of the above-mentioned methods. Wherein, the storage medium may be a magnetic disk, an optical disk, a read only memory (Read Only Memory, ROM) or a random access memory (Random Access Memory, RAM), etc.

The above descriptions are only preferred embodiments of the present invention, and are not intended to limit the present invention. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present invention shall be included in the protection of the present invention. within range.

Claims (9)

1. a bandwidth allocation methods, is characterized in that, comprising:

Controller receives data stream bandwidth solicited message and the data stream priority information of the transmission of each main frame;

According to the described data stream bandwidth solicited message that receives and data stream priority information, and the inbound port bandwidth information of described each main frame and outbound port bandwidth information, generate bandwidth allocation information;

Described bandwidth allocation information is sent to described each main frame.

2. the method for claim 1, it is characterized in that, the data stream bandwidth solicited message received described in described basis and data stream priority information, and the inbound port bandwidth information of described each main frame and outbound port bandwidth information, generate bandwidth allocation information, comprising:

Generate data stream bandwidth request matrix R ⁰and weights matrix W, wherein said R ⁰the i-th row jth column element for main frame M _ito M _jdata flow f _ijbandwidth request, the i-th row jth column element ω of described weights matrix W _ijfor main frame M _ito M _jdata flow f _ijweights, described R ⁰be m × m matrix with W, m represents the number of main frame, i and j is natural number, and described i≤m, j≤m;

To described matrix R ⁰often row, calculated data stream f successively _ijthe bandwidth of according to priority weights distribution wherein $r_{\mathrm{ij}}^{\mathrm{temp}}=R\left({P_i}^{\mathrm{out}}\right)\&times;\frac{{\mathrm{\&omega;}}_{\mathrm{ij}}}{\underset{j=1,...,m}{\mathrm{\&Sigma;}}{\mathrm{\&omega;}}_{\mathrm{ij}}},$ for main frame M _ioutbound port bandwidth information;

Will with smaller value in both is as matrix R ¹in i-th row jth row element

According to described remaining bandwidth, described in renewal

To described matrix R ¹often row, calculated data stream f successively _ijthe bandwidth of according to priority weights distribution wherein $r_{\mathrm{ij}}^{\mathrm{temp}}=R\left({P_i}^{\mathrm{in}}\right)\&times;\frac{{\mathrm{\&omega;}}_{\mathrm{ij}}}{\underset{j=1,...,m}{\mathrm{\&Sigma;}}{\mathrm{\&omega;}}_{\mathrm{ij}}},$ for main frame M _iinbound port bandwidth information;

Will with smaller value in both is as matrix R ²in i-th row jth row element

According to described remaining bandwidth, described in renewal

3. method as claimed in claim 2, is characterized in that, send described bandwidth allocation information, comprising to described each main frame:

The bandwidth allocation information relevant to this main frame is sent to every platform main frame.

4. method as claimed in claim 2, is characterized in that, send described bandwidth allocation information, comprising to described each main frame:

To each main frame sending metrix R ²information, the i-th row jth column element r wherein _ij ²for main frame M _ito M _jdata flow f _ijbandwidth value.

5. a controller, is characterized in that, comprising:

Receiver module, for receiving data stream bandwidth solicited message and the data stream priority information of the transmission of each main frame;

Bandwidth allocation module, for the data stream bandwidth solicited message that receives according to described receiver module and data stream priority information, and the inbound port bandwidth information of described each main frame and outbound port bandwidth information, generate bandwidth allocation information;

Sending module, for sending described bandwidth allocation information to described each main frame.

6. controller as claimed in claim 5, it is characterized in that, described controller also comprises:

Generation module, for generating data stream bandwidth request matrix R ⁰and weights matrix W, wherein said R ⁰the i-th row jth column element for main frame M _ito M _jdata flow f _ijbandwidth request, the i-th row jth column element ω of described weights matrix W _ijfor main frame M _ito M _jdata flow f _ijweights, described R ⁰be m × m matrix with W, m represents the number of main frame, i and j is natural number, and described i≤m, j≤m.

7. controller as claimed in claim 6, is characterized in that, described bandwidth allocation module specifically for:

To described matrix R ⁰often row, calculated data stream f successively _ijthe bandwidth of according to priority weights distribution wherein $r_{\mathrm{ij}}^{\mathrm{temp}}=R\left({P_i}^{\mathrm{out}}\right)\&times;\frac{{\mathrm{\&omega;}}_{\mathrm{ij}}}{\underset{j=1,...,m}{\mathrm{\&Sigma;}}{\mathrm{\&omega;}}_{\mathrm{ij}}},$ for main frame M _ioutbound port bandwidth information;

Will with smaller value in both is as matrix R ¹in i-th row jth row element

According to described remaining bandwidth, described in renewal

To described matrix R ¹often row, calculated data stream f successively _ijthe bandwidth of according to priority weights distribution wherein $r_{\mathrm{ij}}^{\mathrm{temp}}=R\left({P_i}^{\mathrm{in}}\right)\&times;\frac{{\mathrm{\&omega;}}_{\mathrm{ij}}}{\underset{j=1,...,m}{\mathrm{\&Sigma;}}{\mathrm{\&omega;}}_{\mathrm{ij}}},$ for main frame M _iinbound port bandwidth information;

Will with smaller value in both is as matrix R ²in i-th row jth row element

According to described remaining bandwidth, described in renewal

8. controller as claimed in claim 7, is characterized in that, described sending module specifically for:

The bandwidth allocation information relevant to this main frame is sent to every platform main frame; Or to each main frame sending metrix R ²information, the i-th row jth column element wherein for main frame M _ito M _jdata flow f _ijbandwidth value.

9. a communication system, comprise a controller and at least two main frames, it is characterized in that, described controller is used for, and receives data stream bandwidth solicited message and the data stream priority information of the transmission of each main frame;

According to the described data stream bandwidth solicited message that receives and data stream priority information, and the inbound port bandwidth information of described each main frame and outbound port bandwidth information, generate bandwidth allocation information;

Described bandwidth allocation information is sent to described each main frame.