Cloud network service level agreement guarantee method and system based on distributed rate limiting strategy

Abstract

The application discloses a cloud network service level agreement guarantee system and method based on a distributed speed limiting strategy, wherein the distributed speed limiting strategy is realized through an equal packet loss rate control algorithm, and a core purpose of the equal packet loss rate control algorithm is to update a speed limiting value of each gateway node in a distributed speed limiting gateway cluster when each round time is exhausted; the guarantee system comprises a management module and a routing module; the management module is used for starting, monitoring and configuring the distributed speed limiting gateway cluster based on the distributed speed limiting strategy; and the routing module is used for forwarding data packets in a cloud network and performing bandwidth speed limiting based on the distributed speed limiting strategy in the forwarding process. The system and method utilize the distributed speed limiting strategy, realize cloud network service level agreement guarantee, and are favorable for improving cloud network service quality.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of cloud network quality of service guarantee, and particularly relates to a cloud network service level agreement guarantee method and system based on a distributed rate limiting strategy. BACKGROUND

[0002] A decade ago, the global network entered the cloud network era. In recent years, cloud networks have been providing services in all aspects of various industries such as text, video, games, meetings, and public services. For cloud services, bandwidth guarantee with high real-time, high stability, high accuracy, and high reliability is a problem that customers are closely concerned about, and is also an important indicator for measuring the pros and cons of cloud vendor products and services, and all of this depends on the accurate bandwidth limiting capability of the gateway node. The bandwidth limiting capability is an important part of the elastic computing service sold by major cloud service providers, and distributed rate limiting is one of the best choices in the context of the growing size of the cloud network.

[0003] The existing cloud network lacks accurate distributed rate limiting capability, so that when user business traffic changes, network bandwidth will have uncertainty. This uncertainty mainly manifests in that when some traffic changes, network delay, packet loss, or even interruption problems occur, resulting in long non-response of user services and applications, which has a huge impact on user business.

[0004] Rate limiting in a cloud network refers to the bandwidth usage limit for services purchased by users. This limit is to ensure that user services do not affect each other and to ensure that user services do not affect the performance of the network host, thereby not affecting the service itself. In addition, rate limiting means that cloud network service providers need to ensure that they can provide users with rate-limited bandwidth usage.

[0005] Distributed rate limiting in a cloud network refers to rate limiting the same rate limiting granularity service on multiple gateway nodes. Here, multiple gateway nodes usually refer to multiple gateways. As the size of the cloud network increases and user service demand rises, a gateway node often cannot meet the demand, and scenarios requiring multiple gateway nodes to work together require rate limiting information to be synchronized to provide higher real-time, high stability, high accuracy, and high reliability bandwidth guarantee.

[0006] User services often have certain uncertainties, such as uncertainty in traffic type and uncertainty in bandwidth usage. Distributed rate limiting also has certain requirements in terms of communication overhead and bandwidth overhead. Therefore, there has not been a suitable distributed rate limiting strategy.

[0007] However, with the increasing number of users on the cloud, the urgent need for fast response and accurate rate limiting of online elastic computing services has become increasingly prominent, and the contradiction between this and unreliable rate limiting strategies has become more and more prominent. SUMMARY

[0008] In view of the above, the purpose of the present application is to provide a cloud network service level agreement guarantee method and system based on a distributed rate limiting strategy, which realizes cloud network service level agreement guarantee by using a distributed rate limiting strategy, and is conducive to improving the quality of cloud network services.

[0009] To achieve the above-mentioned purpose, the embodiment provides a cloud network service level agreement guarantee system based on a distributed rate limiting strategy, which comprises:

[0010] The distributed rate limiting strategy is realized by an equal packet loss rate control algorithm, and the core purpose of the equal packet loss rate control algorithm is to update the rate limiting value of each gateway node in the distributed rate limiting gateway cluster when each round of time is exhausted.

[0011] The guarantee system comprises a management module and a routing module.

[0012] The management module is configured to start, monitor and configure the distributed rate limiting gateway cluster based on the distributed rate limiting strategy.

[0013] The routing module is configured to forward data packets in the cloud network and perform bandwidth limiting based on the distributed rate limiting strategy during the forwarding process.

[0014] In the embodiment, the equal packet loss rate control algorithm adopts the following formula when updating the rate limiting value of each gateway node:

[0015]

[0016] wherein C i represents the current rate limiting value of the gateway node i, L i represents the next round rate limiting value of the gateway node i, N represents the total number of gateway nodes, d represents the number of communicable gateway nodes of each gateway node, E represents the communicable set of gateway nodes, r i and r j respectively represent the packet receiving rate of the gateway node i and the gateway node j, R i represents the cluster packet receiving rate of the gateway node i, η represents a constant parameter, P i and P j respectively represent the packet loss rate of the gateway node i and the node j.

[0017] In the embodiment, the management module comprises a starting unit, a monitoring unit and a variable unit.

[0018] The starting unit starts the distributed rate limiting gateway cluster and configures the initial parameters of the distributed rate limiting gateway cluster according to the initial parameters configured by the user,

[0019] The monitoring unit is configured to monitor abnormal state information of the distributed rate limiting gateway cluster in runtime, and synchronously monitor normal state information of all gateway nodes in the distributed rate limiting gateway cluster;

[0020] The parameter changing unit is configured to receive changed parameters during runtime of the distributed rate limiting gateway cluster, and send the changed parameters to the distributed rate limiting gateway cluster;

[0021] The initial parameters and the changed parameters both include: total number of gateway nodes, number of communicable gateway nodes of each gateway node, initial rate limiting value of each gateway node, communicable set between gateway nodes, constant parameters, and time information of each round.

[0022] In an embodiment, the abnormal state information includes that a gateway node does not forward data as expected and a gateway node is down; and the normal state information includes a virtual IP to which a network packet belongs, packet receiving rate of the virtual IP, forwarding rate of the virtual IP, and packet loss rate of the virtual IP, and rate limiting value of the virtual IP.

[0023] In an embodiment, the routing module includes a data forwarding unit and a rate limiting unit.

[0024] The data forwarding unit is configured to forward a network packet to a correct network device.

[0025] The rate limiting unit is configured to limit forwarding rate of a network packet according to configuration information, wherein the configuration information includes a first configuration packet sent by the starting unit in starting, a second configuration packet sent by the parameter changing unit in parameter changing, and a third configuration packet sent by the distributed rate limiting gateway cluster in normal operation.

[0026] In an embodiment, the rate limiting unit includes a statistics subunit, a configuration subunit, an analysis subunit, an encapsulation subunit, and a token bucket subunit.

[0027] The statistics subunit is configured to count forwarding data of a gateway node, wherein the counted forwarding data includes a virtual IP to which a network packet belongs, packet receiving rate of the virtual IP, forwarding rate of the virtual IP, and packet loss rate of the virtual IP.

[0028] The encapsulation subunit is configured to encapsulate the counted forwarding data and configuration data of a distributed rate limiting strategy into a third configuration packet.

[0029] The analysis subunit is configured to analyze a packet header of each network packet, distinguish a third configuration packet for configuring a distributed rate limiting strategy, and analyze forwarding data and configuration data contained in the third configuration packet.

[0030] The configuration subunit is configured to update the distributed rate limiting strategy according to data contained in the third configuration packet.

[0031] The token bucket subunit is configured to implement a token bucket algorithm to assist in running a distributed rate limiting strategy.

[0032] In an embodiment, the implementation of the token bucket algorithm to assist in running the distributed rate limiting strategy comprises:

[0033] According to the obtained configuration information, the token upper limit and token generation rate of the primary token bucket and the secondary token bucket are configured for each virtual IP; after starting the routing and forwarding work, the number of tokens in the corresponding token bucket is reduced according to the number of tokens consumed by the data forwarding unit, and when the number of tokens is insufficient for forwarding data packets, the corresponding data packets are discarded.

[0034] To achieve the above-mentioned purposes, the embodiment further provides a cloud network service level agreement guarantee method based on a distributed rate limiting strategy, which adopts the cloud network service level agreement guarantee system and comprises the following steps:

[0035] The distributed rate limiting strategy and the distributed rate limiting gateway cluster are started and configured by the management module;

[0036] The distributed rate limiting strategy is implemented on the distributed rate limiting gateway cluster by the routing module, and network data packets are forwarded and routed according to the distributed rate limiting strategy;

[0037] The distributed rate limiting strategy and the distributed rate limiting gateway cluster are monitored and reconfigured by the management module.

[0038] In an embodiment, the implementation of the distributed rate limiting strategy on the distributed rate limiting gateway cluster by the routing module comprises:

[0039] The configuration subunit initializes the parameters contained in the distributed rate limiting strategy according to the obtained configuration information;

[0040] The token bucket subunit initializes the token bucket, that is, the token upper limit and token generation rate of the primary token bucket and the secondary token bucket are configured for each virtual IP according to the configuration information;

[0041] When the distributed rate limiting gateway cluster is running, the analysis subunit distinguishes between normal data packets and third configuration data packets related to the distributed rate limiting strategy, the data forwarding unit forwards the normal data packets and the token bucket subunit limits the forwarding of the normal data packets, and each forwarded or discarded data packet is counted by the counting subunit;

[0042] According to the per-round time information in the distributed rate limiting strategy configuration information, the information of the current round is encapsulated into the third configuration data packet by the counting subunit and the encapsulation subunit at the end of each round, and then forwarded to the corresponding network device by the data forwarding unit;

[0043] The distributed rate limiting strategy related data contained in the third configuration data packet is parsed by a parsing subunit, and then the distributed rate limiting strategy related parameters are updated by a configuration subunit.

[0044] In the embodiment, the distributed rate limiting strategy and the distributed rate limiting gateway cluster are monitored by the management module, including:

[0045] The monitoring unit constructs a detection data packet and sends it to all gateway nodes in the distributed rate limiting gateway cluster, collects the response data packets returned by the gateway nodes within a certain window time, analyzes the response data packets and judges the gateway node state information, and according to the analysis and judgment result, encapsulates the normal state and configuration information into a network data packet and sends it to the gateway nodes in normal state again.

[0046] According to the normal state information of the monitoring unit and the artificially configured change parameters, the change parameters are sent to the gateway nodes in normal state in the distributed rate limiting gateway cluster through the variable configuration unit.

[0047] Compared with the prior art, the present application has at least the following beneficial effects:

[0048] (1) In the cloud network, multiple gateway nodes can be used to limit the data packets sent to the same virtual IP, improving the upper limit of the available bandwidth of a single virtual IP and reducing the occurrence of delay increase and disconnection problems;

[0049] (2) The dynamic adjustment of the rate limiting value is realized on the cloud network multi-node cluster distributed rate limiting strategy, improving the rate limiting accuracy and precision, speeding up the convergence of the gateway node cluster to the balanced state during variable configuration, and improving the service quality, wherein the balanced state refers to the consistent packet loss rate of each node mentioned in the algorithm;

[0050] (3) The function of updating the rate limiting parameters through mutual communication of the nodes is added to the cloud network multi-node cluster distributed rate limiting strategy, reducing the bandwidth overhead of communication. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the following embodiment or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0052] Figure 1 is a structural schematic diagram of the cloud network service level agreement guarantee system based on the distributed rate limiting strategy provided by the embodiment;

[0053] Figure 2 is a structural schematic diagram of the management module provided by the embodiment;

[0054] Figure 3 is a structural schematic diagram of the routing module provided by the embodiment;

[0055] Figure 4 is a structural schematic diagram of the rate limiting unit provided by the embodiment;

[0056] Figure 5 is a flowchart of the cloud network service level agreement guarantee method based on the distributed rate limiting strategy provided by the embodiment;

[0057] Figure 6 is a working flowchart of the routing module provided by the embodiment;

[0058] Figure 7 is a working flowchart of the management module monitoring and reconfiguring the distributed rate limiting strategy and the distributed rate limiting gateway cluster provided by the embodiment. DETAILED DESCRIPTION

[0059] To make the objectives, technical solutions and advantages of the present application clearer, further detailed description will be given below in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the protection scope of the present application.

[0060] Figure 1 is a structural and working flowchart of the cloud network service level agreement guarantee system based on the distributed rate limiting strategy provided by the embodiment, as shown in Figure 1 The cloud network service level agreement guarantee system based on the distributed rate limiting strategy provided by the embodiment includes a management module and a routing module to realize starting, monitoring and reconfiguring the distributed rate limiting gateway cluster and maintaining the normal operation of the distributed rate limiting strategy, wherein the client is used as the main body of the sold elastic network service.

[0061] The distributed rate limiting strategy is realized through the equal packet loss rate control algorithm, which can guarantee the stability of the distributed rate limiting gateway cluster and realize the distributed rate limiting. The specific parameter index is that the packet loss rate of each gateway node of the distributed rate limiting gateway cluster tends to be consistent. In addition to the packet loss rate P i of each gateway node of the distributed rate limiting gateway cluster, the parameters involved in the equal packet loss rate control algorithm also include: the current rate limiting value C i of node i, the next round rate limiting value L i of node i, the total number of nodes N, the number of communicable nodes d of each node, the inter-node communicable set E, the packet receiving rate r i of node i, the cluster packet receiving rate R i of node i and the constant parameter η. Among them, C i and L iis the speed limit value in the distributed speed limit strategy, d is the number of nodes that each gateway node can directly communicate with (i.e. the number of nodes that a node can reach in one hop in the virtual network topology), the set E contains a plurality of pairs of numbers (i, j) representing that the node i and the node j can directly communicate, R i =∑r j +r i , (i, j) E; the constant parameter η controls the variation range of the speed limit value in each round.

[0062] Among the above parameters, C i , N, d, E, η can be configured by the starting unit and the variable unit, P i , L i , r i , R i are obtained by the statistical subunit and the analysis subunit. It should be noted that the calculation of some parameters such as P i , C i , η, L i , r i , R i is related to the time interval of each round, and the time interval of each round is also configured by the starting unit and the variable unit. The core purpose of the packet loss rate control algorithm is to update the speed limit value of each gateway node in the distributed speed limit gateway cluster when each round of time is consumed using the following formula:

[0063]

[0064] The management module is used to start, monitor and configure the distributed speed limit gateway cluster based on the distributed speed limit strategy. Specifically, the elastic computing service control initiates a start instruction to indicate that the distributed speed limit strategy starts to be applied to the distributed speed limit gateway cluster. After receiving the distributed speed limit strategy start instruction, the management module applies the distributed speed limit strategy to the distributed speed limit gateway cluster, including starting and configuring the gateway cluster, detecting the gateway cluster, and adjusting the gateway cluster. The elastic computing service is a cloud network service sold by a network vendor, which is essentially a virtual machine. The elastic computing service control will apply the distributed speed limit strategy according to the actual situation due to the load balancing of the vendor's network computer cluster or the customer's demand.

[0065] As shown in Figure 2 , the management module includes a starting unit, a monitoring unit and a variable unit. The starting unit starts the distributed speed limit gateway cluster and configures the initial parameters of the distributed speed limit gateway cluster according to the initial parameters configured by the user. Generally, the gateway is the uppermost core node of the cloud network, and has strong traffic forwarding and routing capabilities. Only when the single gateway node capacity does not meet the business demand, the distributed speed limit strategy is applied.

[0066] The monitoring unit is used for monitoring abnormal state information of the distributed rate limiting gateway cluster in runtime, and also synchronizing normal state information of all gateway nodes in the distributed rate limiting gateway cluster. The abnormal state information includes that the gateway node does not forward the rate limiting data as expected and the gateway node is down. The gateway node down is caused by code bug, machine failure, local power failure and other problems. The normal state information includes a virtual IP to which a data packet belongs, a packet receiving rate of the virtual IP, a forwarding rate of the virtual IP, a packet loss rate of the virtual IP and a rate limiting value of the virtual IP. The specific monitoring process includes:

[0067] The monitoring unit constructs a probe data packet and sends it to all gateway nodes in the distributed rate limiting gateway cluster, collects response data packets returned by the gateway nodes in a certain window time, analyzes the response data packets and judges the state information of the gateway nodes, encapsulates the normal state and configuration information into a network data packet according to the analysis and judgment results, and sends it to the gateway nodes that are not down again, reconstructs the probe data packet and sends it after a certain time interval, and repeats the process.

[0068] The configuration unit is used for receiving changed parameters configured during the running of the distributed rate limiting gateway cluster, and distributing the changed parameters to the distributed rate limiting gateway cluster. The initial parameters and the changed parameters both include C i , N, d, E, η and time interval information of each round. The specific configuration process includes receiving the changed parameters input by the user, constructing a second configuration data packet based on the changed parameters and sending it to the corresponding gateway node, judging whether the configuration is successful according to the state information collected by the monitoring unit, ending the configuration process if the configuration is successful, or continuing the configuration by inputting the changed parameters again.

[0069] The routing module is used for forwarding and routing data packets in the cloud network, and performing bandwidth rate limiting based on the distributed rate limiting strategy during the forwarding process. Specifically, it includes starting, configuring the distributed rate limiting strategy and the distributed rate limiting gateway cluster in cooperation with the management module; implementing the distributed rate limiting strategy on the distributed rate limiting gateway cluster, and correctly forwarding and routing the network data packets according to the distributed rate limiting strategy; monitoring and reconfiguring the distributed rate limiting strategy and the distributed rate limiting gateway cluster in cooperation with the management module. It should be noted that the configuration data packet is a kind of network data packet.

[0070] As shown in Figure 3 , the routing module includes a data forwarding unit and a rate limiting unit. The data forwarding unit is used for forwarding network data packets to correct network devices. In the cloud network, the forwarding of network data packets is similar to the physical network, the difference being that the forwarding devices of the cloud network are mostly software virtual or software and hardware combined virtual switches, etc., and the IP addresses used are mostly virtual IPs, which are converted by the virtual IP-physical IP conversion table, so as to ensure that they can be sent to the correct physical devices.

[0071] The rate limiting unit is configured to limit the forwarding rate of the data packet according to the configuration information, wherein the configuration information includes the first configuration data packet sent by the starting unit when starting, the second configuration data packet sent by the reconfiguring unit when reconfiguring, and the third configuration data packet sent by the distributed rate limiting gateway cluster when normally operating and self-adjusting.

[0072] As shown in Figure 4 The rate limiting unit includes a statistics subunit, a configuration subunit, an analysis subunit, an encapsulation subunit, and a token bucket subunit. The statistics subunit is configured to count the forwarding data of the gateway node, wherein the counted forwarding data includes the virtual IP to which the network data packet belongs, the packet receiving rate of the virtual IP, the forwarding rate of the virtual IP, and the packet loss rate of the virtual IP. The encapsulation subunit is configured to encapsulate the counted forwarding information and the configuration data of the distributed rate limiting strategy into the third configuration data packet. The analysis subunit is configured to analyze the packet header of each network data packet, distinguish the third configuration data packet of the distributed rate limiting strategy, and analyze the data contained in the third configuration data packet. The configuration subunit is configured to update the distributed rate limiting strategy according to the conversion data and the configuration data contained in the third configuration data packet. The token bucket subunit is configured to implement the token bucket algorithm to assist in running the distributed rate limiting strategy.

[0073] In the token bucket subunit, implementing the token bucket algorithm to assist in running the distributed rate limiting strategy includes: configuring the upper limit of the tokens of the primary token bucket and the secondary token bucket and the token generation rate for each virtual IP according to the obtained configuration information; after starting the routing and forwarding work, reducing the number of tokens in the corresponding token bucket according to the number of tokens consumed by the data forwarding unit; and when the number of tokens is insufficient for forwarding the data packet, discarding the corresponding data packet. All data in the process of implementing the token bucket algorithm to assist in running the distributed rate limiting strategy are counted by the statistics subunit and encapsulated into the third configuration data packet by the encapsulation subunit at the end of each round of time.

[0074] In the embodiment, based on the cloud network service level agreement guarantee system, the embodiment further provides a cloud network service level agreement guarantee method based on a distributed rate limiting strategy, as shown in Figure 5 The method includes step 1 and parallel steps 2 and 3.

[0075] In step 1, the distributed rate limiting strategy and the distributed rate limiting gateway cluster are started and configured by the management module. Specifically, the distributed rate limiting gateway cluster is started and the initial parameters are configured by the starting unit.

[0076] In step 2, the distributed rate limiting strategy is implemented on the distributed rate limiting gateway cluster by the routing module, and the network data packet is forwarded and routed according to the distributed rate limiting strategy. As shown in Figure 6 The specific process is as follows:

[0077] Step 2-1, initializing the parameters contained in the distributed rate limiting strategy according to the obtained configuration information by the configuration subunit;

[0078] Step 2-2, initializing the token bucket by the token bucket subunit, i.e. configuring the upper limit of tokens and the token generation rate of the primary token bucket and the secondary token bucket for each virtual IP according to the configuration information;

[0079] Step 2-3, starting the distributed rate limiting gateway cluster, distinguishing normal data packets from the third configuration data packets related to the distributed rate limiting strategy by the parsing subunit, forwarding the normal data packets by the data forwarding unit and limiting the forwarding of the normal data packets by the token bucket subunit, and counting each forwarded or dropped data packet by the counting subunit;

[0080] Step 2-4, according to the time information of each round in the distributed rate limiting strategy configuration information, encapsulating the information of the current round into the third configuration data packet by the counting subunit and the encapsulation subunit at the end of each round, and then forwarding it to the corresponding network device by the data forwarding unit;

[0081] Step 2-5, parsing the distributed rate limiting strategy related data contained in the third configuration data packet by the parsing subunit, and then updating the distributed rate limiting strategy related parameters by the configuration subunit.

[0082] It should be noted that steps 2-3 to 2-5 are repeated during the operation of the distributed rate limiting gateway cluster, but the process of starting the distributed rate limiting gateway cluster is not included.

[0083] In step 3, the distributed rate limiting strategy and the distributed rate limiting gateway cluster are monitored and reconfigured by the management module. Specifically, (a) the running status of the distributed rate limiting gateway cluster is monitored by the monitoring unit, and if an abnormal condition is found, it is automatically handled according to the preconfigured scheme, such as Figure 7 As shown, the specific monitoring process is: the monitoring unit constructs a probe data packet and sends it to all gateway nodes in the distributed rate limiting gateway cluster, collects the response data packets returned by the gateway nodes within a certain window time, and judges the gateway node state information based on the response data packets, encapsulates the normal state and configuration information into a network data packet and sends it to the gateway nodes that are not down, after a certain time interval, the probe data packet is reconstructed and sent again, and the process is repeated; (b) according to the normal state information of the monitoring unit and the manually configured change parameters, the change parameters are sent to the gateway nodes in the normal state of the distributed rate limiting gateway cluster by the variable configuration unit. It should be noted that steps (a) and (b) are also repeated and alternated processes.

[0084] The cloud network service level agreement guarantee system and method based on the distributed speed limiting strategy provided by the above embodiment realize cloud network level agreement guarantee based on the distributed speed limiting strategy, and are beneficial to improving the service quality of the cloud network.

[0085] The above detailed description of the specific embodiments has described the technical solutions and beneficial effects of the present application. It should be understood that the above description is only the most preferred embodiment of the present application and is not used to limit the present application. Any modification, supplement and equivalent replacement made within the principle range of the present application should be included in the protection range of the present application.

Claims

1. A cloud network service level protocol guarantee system based on a distributed rate limiting strategy, characterized in that, The distributed rate limiting strategy is implemented through an equal packet loss rate control algorithm. The core purpose of this algorithm is to update the rate limiting value of each gateway node in the distributed rate limiting gateway cluster when the time of each round expires. The protection system includes a management module and a routing module; The management module is used to start, monitor, and configure the distributed rate limiting gateway cluster based on the distributed rate limiting policy. The routing module is used to forward data packets in the routing cloud network and to perform bandwidth limiting based on a distributed rate limiting strategy during the forwarding process. The equal packet loss rate control algorithm uses the following formula when updating the rate limit value of each gateway node: in, Indicates gateway node i The current speed limit value, Indicates gateway node i The next speed limit value, This indicates the total number of gateway nodes. This indicates the number of communicable gateway nodes for each gateway node. This represents the set of nodes that can communicate with each other. and Representing gateway nodes i and gateway node j Packet reception rate, Indicates gateway node i Cluster packet reception rate, Represents a constant parameter. and Representing gateway nodes i and gateway node j The packet loss rate.

2. The cloud network service level protocol guarantee system based on distributed rate limiting strategy according to claim 1, characterized in that, The management module includes a startup unit, a monitoring unit, and a configuration change unit; The startup unit starts the distributed rate-limiting gateway cluster and configures the initial parameters of the distributed rate-limiting gateway cluster according to the initial parameters configured by the user. The monitoring unit is used to monitor abnormal status information during the operation of the distributed rate limiting gateway cluster, and also to synchronize the normal status information of all gateway nodes in the distributed rate limiting gateway cluster. The configuration unit is used to receive the configuration modification parameters during the operation of the distributed rate limiting gateway cluster and send the modification parameters to the distributed rate limiting gateway cluster. Both the initial and modified parameters include: the total number of gateway nodes, the number of communicable gateway nodes for each gateway node, the initial speed limit for each gateway node, the set of communicable gateway nodes, constant parameters, and time information for each round.

3. The cloud network service level protocol guarantee system based on distributed rate limiting strategy according to claim 2, characterized in that, The abnormal status information includes the gateway node failing to forward and rate-limit data as expected, and the gateway node crashing; the normal status information includes the virtual IP to which the data packet belongs, the packet receiving rate of the virtual IP, the forwarding rate of the virtual IP, the packet loss rate of the virtual IP, and the rate limit value of the virtual IP.

4. The cloud network service level protocol guarantee system based on distributed rate limiting strategy according to claim 1, characterized in that, The routing module includes a data forwarding unit and a rate limiting unit; The data forwarding unit is used to forward network data packets to the correct network device; The rate limiting unit is used to limit the forwarding rate of data packets according to the configuration information, which includes a first configuration data packet sent by the startup unit at startup, a second configuration data packet sent by the configuration change unit at change time, and a third configuration data packet sent by the distributed rate limiting gateway cluster during normal operation.

5. The cloud network service level protocol guarantee system based on distributed rate limiting strategy according to claim 4, characterized in that, The rate limiting unit includes a statistics subunit, a configuration subunit, a parsing subunit, an encapsulation subunit, and a token bucket unit; The statistics subunit is used to count the forwarding data of the gateway node. The counted forwarding data includes the virtual IP to which the network data packet belongs, the packet receiving rate of the virtual IP, the forwarding rate of the virtual IP, and the packet loss rate of the virtual IP. The encapsulation subunit is used to encapsulate the statistical forwarding data and the configuration data of the distributed rate limiting policy into a third configuration data packet; The parsing subunit is used to analyze the header of each network data packet, find the third configuration data packet that configures the distributed rate limiting policy, and parse the forwarding data and configuration data contained in the third configuration data packet; The configuration subunit is used to update the distributed rate limiting policy based on the data contained in the third configuration data packet; The token bucket subunit is used to implement the token bucket algorithm to assist in the operation of the distributed rate limiting strategy.

6. The cloud network service level protocol guarantee system based on distributed rate limiting strategy according to claim 5, characterized in that, The implementation of the token bucket algorithm to assist in the operation of the distributed rate limiting strategy includes: Based on the obtained configuration information, configure the token limit and token generation rate of the primary and secondary token buckets for each virtual IP. After the routing and forwarding work starts, reduce the number of tokens in the corresponding token bucket according to the number of tokens required by the data forwarding unit. When the number of tokens is insufficient to forward data packets, discard the corresponding data packets.

7. A cloud network service level protocol guarantee method based on a distributed rate limiting strategy, characterized in that, The method employs the cloud network service level protocol protection system as described in any one of claims 1-6, and includes the following steps: Start and configure the distributed rate limiting policy and distributed rate limiting gateway cluster through the management module; The routing module implements distributed rate limiting policies on the distributed rate limiting gateway cluster and forwards and routes network data packets according to the distributed rate limiting policies. The management module monitors and reconfigures distributed rate limiting policies and distributed rate limiting gateway clusters.

8. The cloud network service level protocol guarantee method based on distributed rate limiting strategy according to claim 7, characterized in that, The implementation of distributed rate limiting policies on the distributed rate limiting gateway cluster via the routing module includes: The configuration subunit initializes the parameters included in the distributed rate limiting policy based on the acquired configuration information; The token bucket is initialized through the token bucket subunit, that is, the token limit and token generation rate of the primary and secondary token buckets are configured for each virtual IP according to the configuration information. When the distributed rate limiting gateway cluster is running, the parsing subunit distinguishes normal data packets from third configuration data packets related to the distributed rate limiting policy. The data forwarding unit forwards normal data packets and the token bucket subunit restricts the forwarding of normal data packets. Each forwarded or lost data packet should be statistically analyzed by the statistics subunit. Based on the time information of each round in the distributed rate limiting policy configuration information, at the end of each round, the information of this round is encapsulated into the third configuration data packet by the statistics subunit and the encapsulation subunit, and then forwarded to the corresponding network device by the data forwarding unit; The distributed rate limiting policy-related data contained in the third configuration data packet is parsed by the parsing subunit, and then the distributed rate limiting policy-related parameters are updated by the configuration subunit.

9. The cloud network service level protocol guarantee method based on distributed rate limiting strategy according to claim 7, characterized in that, The management module monitors distributed rate-limiting policies and the distributed rate-limiting gateway cluster, including: The monitoring unit constructs probe data packets and sends them to all gateway nodes in the distributed rate-limiting gateway cluster. Within a certain window time, it collects the response data packets returned by the gateway nodes, analyzes the response data packets and judges the status information of the gateway nodes. Based on the analysis and judgment results, it encapsulates the normal status information and configuration information into network data packets and sends them again to the gateway nodes that have not crashed. Based on the normal status information of the monitoring unit and the manually configured modified parameters, the modified parameters are sent to the gateway nodes in the distributed rate limiting gateway cluster that are in normal status through the configuration change unit.

Images