WO2021004063A1 - 一种缓存服务器的带宽调度方法及装置 - Google Patents

一种缓存服务器的带宽调度方法及装置 Download PDF

Info

Publication number
WO2021004063A1
WO2021004063A1 PCT/CN2020/074968 CN2020074968W WO2021004063A1 WO 2021004063 A1 WO2021004063 A1 WO 2021004063A1 CN 2020074968 W CN2020074968 W CN 2020074968W WO 2021004063 A1 WO2021004063 A1 WO 2021004063A1
Authority
WO
WIPO (PCT)
Prior art keywords
target
health value
cache server
bandwidth
target cache
Prior art date
Application number
PCT/CN2020/074968
Other languages
English (en)
French (fr)
Inventor
向校民
Original Assignee
网宿科技股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 网宿科技股份有限公司 filed Critical 网宿科技股份有限公司
Priority to EP20775568.7A priority Critical patent/EP3783869A4/en
Priority to US17/060,032 priority patent/US11102286B2/en
Publication of WO2021004063A1 publication Critical patent/WO2021004063A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004—Server selection for load balancing
    • H04L67/101—Server selection for load balancing based on network conditions
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004—Server selection for load balancing
    • H04L67/1023—Server selection for load balancing based on a hash applied to IP addresses or costs
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004—Server selection for load balancing
    • H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004—Server selection for load balancing
    • H04L67/1012—Server selection for load balancing based on compliance of requirements or conditions with available server resources
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1029—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers using data related to the state of servers by a load balancer
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/50—Network services
    • H04L67/56—Provisioning of proxy services
    • H04L67/568—Storing data temporarily at an intermediate stage, e.g. caching
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/70—Admission control; Resource allocation
    • H04L47/80—Actions related to the user profile or the type of traffic
    • H04L47/803—Application aware
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/70—Admission control; Resource allocation
    • H04L47/80—Actions related to the user profile or the type of traffic
    • H04L47/805—QOS or priority aware

Definitions

  • This application relates to the field of Internet technology, and in particular to a bandwidth scheduling method and device of a cache server.
  • CDN Content Delivery Network
  • CDN service providers can set the rated bandwidth for each cache server deployed, and maintain the real-time bandwidth of the cache server through bandwidth scheduling Below the rated bandwidth.
  • the technical staff of the CDN service provider can collect the real-time bandwidth of each cache server in real time through the back-end server, and calculate the bandwidth call amount according to the real-time bandwidth and rated bandwidth of each cache server, that is, need to be dispatched to other cache servers , The amount of bandwidth served by other cache servers, or the amount of bandwidth transferred in is calculated, that is, the amount of bandwidth that the cache server can carry, and bandwidth scheduling is performed according to the amount of bandwidth transferred out/in.
  • the background server collects that the real-time bandwidth of a cache server is 800 MB. Assuming that the rated bandwidth of the cache server is 700 MB, the background server can conclude that the bandwidth of the cache server is 100 MB; assuming that the cache server has a bandwidth of 100 MB. The rated bandwidth is 1000 megabytes, and the background server can determine that the bandwidth of the cache server is 200 megabytes.
  • the load capacity of the cache server is not only affected by the total bandwidth carried, but also affected by factors such as the type of service requested, the number of connections, etc., it may happen that after bandwidth scheduling is performed on a cache server according to the amount of bandwidth call, the cache The server load is still high, or after bandwidth scheduling is performed on a cache server according to the amount of bandwidth transferred, the cache server load is too high, which affects the network service quality.
  • embodiments of the present application provide a bandwidth scheduling method and device for a cache server.
  • the technical solution is as follows.
  • a bandwidth scheduling method of a cache server includes:
  • the method further includes:
  • the updating the target health value calculation formula of the target cache server based on the historical health value and the historical visit volume information includes:
  • the performing bandwidth scheduling on the target cache server based on the latest visit volume information, the target health value calculation formula, and the rated health value includes:
  • the client When the health value of the target cache server is less than or equal to the rated health value, the client is called for bandwidth scheduling according to the target.
  • the target health value of the target cache server after calculating the target health value of the target cache server based on the target health value calculation formula and the latest visit volume information, and determining whether the target health value is greater than the rated health value, Also includes:
  • the health value of the target cache server is less than or equal to the rated health value, and the difference between the rated health value and the health value of the target cache server is less than or equal to a preset difference value, adjust according to the target Incoming customers for bandwidth scheduling.
  • determining the call-out client of a cache server other than the target cache server at the current moment includes:
  • the target health value is less than the warning health value, it is determined that the call-out clients of other cache servers except the target cache server at the current moment.
  • the performing bandwidth scheduling on the target cache server based on the latest visit volume information, the target health value calculation formula, and the rated health value includes:
  • the reduced access volume information of the target service type is determined as the target call-out volume, and bandwidth scheduling is performed according to the target call-out volume.
  • the target health value of the target cache server after calculating the target health value of the target cache server based on the target health value calculation formula and the latest visit volume information, and determining whether the target health value is greater than the rated health value, Also includes:
  • the The increased access volume information is determined as the target transfer volume, and bandwidth scheduling is performed according to the target transfer volume.
  • the performing bandwidth scheduling on the target cache server based on the latest visit volume information, the target health value calculation formula, and the rated health value includes:
  • the target health value of the target cache server after calculating the target health value of the target cache server based on the target health value calculation formula and the latest visit volume information, and determining whether the target health value is greater than the rated health value, Also includes:
  • the traffic information under each service type includes at least live broadcast bandwidth, live broadcast connections, on-demand bandwidth, on-demand connections, webpage bandwidth, and webpage connections.
  • a bandwidth scheduling device of a cache server includes an information acquisition module and a scheduling module, wherein:
  • the information acquisition module is configured to acquire the latest traffic information of each service type of the target cache server at the beginning of the cycle according to a preset detection cycle;
  • the scheduling module is configured to perform bandwidth scheduling on the target cache server based on the latest visit volume information, the target health value calculation formula, and the rated health value.
  • the device further includes a calculation module and a formula update module, wherein:
  • the calculation module is configured to obtain the statistical data of each performance index of the target cache server in the historical time period before the cycle start time according to the preset formula update period, and calculate the target based on the statistical data and the preset performance health formula The historical health value of the cache server at each time within the historical duration;
  • the formula update module is used to obtain historical visit volume information of each service type in the historical time period before the cycle start time of the target cache server, and based on the historical health value and the historical visit volume The information updates the target health value calculation formula of the target cache server.
  • the formula update module is further used for:
  • the device further includes a judging module for:
  • the scheduling module is also used for:
  • the client When the health value of the target cache server is less than or equal to the rated health value, the client is called for bandwidth scheduling according to the target.
  • the scheduling module is further used for:
  • the health value of the target cache server is less than or equal to the rated health value, and the difference between the rated health value and the health value of the target cache server is less than or equal to a preset difference value, adjust according to the target Incoming customers for bandwidth scheduling.
  • the scheduling module is further used for:
  • the judgment module is used to:
  • the scheduling module is also used for:
  • the reduced access volume information of the target service type is determined as the target call-out volume, and bandwidth scheduling is performed according to the target call-out volume.
  • the scheduling module is further used for:
  • the The increased access volume information is determined as the target transfer volume, and bandwidth scheduling is performed according to the target transfer volume.
  • the judgment module is used to:
  • the scheduling module is also used for:
  • the scheduling module is further used for:
  • the traffic information under each service type includes at least live broadcast bandwidth, live broadcast connections, on-demand bandwidth, on-demand connections, webpage bandwidth, and webpage connections.
  • a server in a third aspect, includes a processor and a memory, the memory stores at least one instruction, at least one program, a code set, or an instruction set, the at least one instruction, the at least one program The code set or instruction set is loaded and executed by the processor to implement the bandwidth scheduling method of the cache server as described in the first aspect.
  • a computer-readable storage medium stores at least one instruction, at least one program, code set, or instruction set, the at least one instruction, the at least one program, the code
  • the set or instruction set is loaded and executed by the processor to implement the bandwidth scheduling method of the cache server as described in the first aspect.
  • the technical solution provided by the embodiment of the application brings the beneficial effects that: in this embodiment, the latest traffic information of the target cache server under each service type at the beginning of the cycle is obtained according to the preset detection period; The target health value calculation formula fitted by the historical health value of the target cache server and the historical visit volume information under each service type, and the rated health value of the target cache server; based on the latest visit volume information and the target health The value calculation formula and the rated health value perform bandwidth scheduling on the target cache server.
  • the back-end server can determine the health status of the cache server based on the access volume information, health value calculation formula and rated health value of the cache server under each service type, and can perform bandwidth scheduling on the cache server according to the access volume information under each service type , In order to keep the load condition of the cache server in good condition, which can effectively improve the network service quality.
  • FIG. 1 is a flowchart of a bandwidth scheduling method of a cache server according to an embodiment of the present application
  • FIG. 2 is a schematic structural diagram of a bandwidth scheduling device of a cache server according to an embodiment of the present application
  • FIG. 3 is a schematic structural diagram of a bandwidth scheduling device of a cache server according to an embodiment of the present application
  • Fig. 4 is a schematic structural diagram of a server provided by an embodiment of the present application.
  • the embodiment of the application provides a bandwidth scheduling method for a cache server.
  • the execution body of the method can be a backend server, which can be any server, or a server cluster composed of multiple servers, and the backend server can be deployed In the computer room of the CDN service provider.
  • the back-end server can perform bandwidth scheduling on the cache server according to the traffic information of each service type according to the calculation formula of the health value of the cache server, the rated health value of the cache server, and the latest access information of the cache server.
  • the above-mentioned background server may include a processor, a memory, and a transceiver.
  • the processor may be used to perform bandwidth scheduling processing for the cache server in the following process, and the memory may be used to store the data required and generated during the following processing.
  • the transceiver can be used to receive and send related data in the following processing.
  • Step 101 Obtain the latest traffic information of each service type of the target cache server at the beginning of the cycle according to a preset detection cycle.
  • CDN service providers can provide customers with network acceleration services through cache servers deployed in various regions.
  • each cache server When each cache server provides customers with network acceleration services, it can record its own performance indicators (such as CPU utilization, memory utilization) , Connection response time, hard disk utilization, etc.), and generated traffic information under each service type.
  • the traffic information under each service type can at least include live broadcast bandwidth, live connection number, on-demand bandwidth, on-demand Number of connections, page bandwidth, number of page connections, etc. Considering that the amount of bandwidth and the number of connections of different business types may have varying degrees of impact on the load status of the cache server, when the cache server load is too high, it will affect the network service quality of the cache server.
  • the bandwidth of the cache server can be scheduled through the background server with the granularity of each service type and the number of connections corresponding to each service type.
  • the background server can obtain the latest traffic information of each service type that a certain cache server (which can be called a target cache server) at the beginning of the current cycle according to a preset detection cycle, for example, at the beginning of the current cycle
  • the latest traffic information such as live bandwidth, live connections, on-demand bandwidth, on-demand connections, web bandwidth, and web connections that have been generated at any moment. So that the background server can perform bandwidth scheduling on the target cache server based on the latest traffic information, the health value calculation formula of the target cache server (which can be called the target health value calculation formula), and the rated health value according to the traffic information under each service type. .
  • Step 102 Obtain a target health value calculation formula fitted based on the historical health value of the target cache server and historical visit volume information under each service type, and the rated health value of the target cache server.
  • the back-end server can fit a health value calculation formula for each cache server based on the historical health value of each cache server and the historical traffic information of each cache server under each business type, and provide each cache server with a health value calculation formula.
  • the server sets a rated health value so that the backend server can calculate the health value of the cache server through the health value calculation formula, and determine the load status of the cache server according to the health value and the rated health value.
  • the health value of the cache server is greater than the rated health Value, the cache server is determined to be unhealthy (that is, the load is too high), otherwise, the cache server is determined to be healthy.
  • the background server After the background server obtains the latest traffic information of the target cache server under each business type at the beginning of the cycle, it can obtain the target fitted based on the historical health value of the target cache server and the historical traffic information under each business type The calculation formula of the target health value of the cache server and the rated health value of the target cache server.
  • the target cache server may update the target health value calculation formula according to the preset formula update period, and the corresponding processing may be as follows: According to the preset formula update period, obtain the target cache server's historical duration before the cycle start time The statistical data of each performance index is calculated based on the statistical data and the preset performance health formula to calculate the historical health value of the target cache server at each time in the historical period. Obtain the historical visit volume information of each service type of the target cache server in the historical time period before the cycle start time, and update the target health value calculation formula of the target cache server based on the historical health value and historical visit volume information.
  • the historical traffic information under each service type includes at least live broadcast bandwidth, live broadcast connections, on-demand bandwidth, on-demand connections, webpage bandwidth, and webpage connections.
  • the technical staff of the CDN service provider can preset a health value for each cache server in the background server with the above performance indicators (such as CPU utilization, memory utilization, connection response time, hard disk utilization, etc.) as variables
  • the calculation formula (which can be called the preset performance health formula)
  • the background server can obtain the historical duration of the target cache server before the start of the current formula update cycle according to the preset formula update cycle (for example, it can be 1 hour)
  • the historical health value of the target cache server at each time in the above historical time period is calculated (for example, the target can be calculated at a granularity of 1 minute).
  • the backend server After the backend server calculates the historical health value of the target cache server at each time in the above historical time period, it can obtain the historical visit volume information of each service type of the target cache server in the historical time period before the start of the above cycle, and based The historical visit volume information calculates the historical visit volume information at each time corresponding to the above-mentioned historical health values. Then, the back-end server can use the regression analysis model in machine learning (such as linear regression, polynomial regression, etc.) based on the statistics of the historical traffic information under each business type corresponding to each time and the historical health value corresponding to each time mentioned above. The health value calculation formula of the target cache server is combined, and the health value calculation formula is updated to the target health value calculation formula of the target cache server.
  • machine learning such as linear regression, polynomial regression, etc.
  • the backend server may also update the target health value calculation formula in combination with the server category of the target cache server.
  • the corresponding processing may be as follows: determine the target server category to which the target cache server belongs, and based on the historical health value, target server category, and The historical traffic information updates the target health value calculation formula of the target cache server.
  • the technical staff of the CDN service provider can use the background server to follow the cache server
  • the hardware configuration of the cache server is divided into different server categories. For example, cache servers with the same hardware configuration such as CPU, memory, and hard disk can be divided into the same server category.
  • the background server can determine the server category to which the target cache server belongs (which can be referred to as the server category) after obtaining the historical traffic information of the target cache server in the historical time period before the cycle start time of the formula update cycle.
  • Target server category based on the target server category, the historical health value of the target cache server at each time in the historical period, and the historical visit volume of each service type of the target cache server in the historical period before the start of the cycle Information, update the target health value calculation formula of the target cache server.
  • Step 103 Perform bandwidth scheduling on the target cache server based on the latest traffic information, the target health value calculation formula, and the rated health value.
  • the background server after the background server obtains the latest visit volume information, target health value calculation formula, and rated health value of the target cache server, it can calculate based on the latest visit volume information of the target cache server and the target health value of the target cache server.
  • the formula calculates the health value of the target cache server.
  • the background server may perform bandwidth scheduling on the target cache server according to the health value of the target cache server, the aforementioned rated health value of the target cache server, the target health value calculation formula, and the above-mentioned latest traffic information.
  • the health value of the server x 1 represents the current live bandwidth of the target cache server, x 2 represents the on-demand bandwidth of the target cache server at the current time, x 3 represents the web bandwidth of the target cache server at the current time, and x 4 represents the current time of the target cache server Number of live broadcast connections, x 5 represents the number of on-demand connections to the target cache server at the current moment, x 6 represents the number of web page connections to the target cache server at the current moment, a 1 , a 2 , a 3 , b 1 , b 2 , and b 3 are each The coefficient corresponding to the traffic information under the service type.
  • the latest traffic information of the target cache server at 10:00 under each service type is 800M live bandwidth, 700M bandwidth on demand, 900M web bandwidth, and 20 live connections.
  • the number of on-demand connections is 15, the number of webpage connections is 40, a 1 , a 2 , a 3 , b 1 , b 2 , and b 3 are 0.03, 0.04, 0.02, 0.35, 0.2, 0.1, respectively.
  • the backend server can be based on the above target health
  • the background server may call the target cache server bandwidth when the target health value of the target cache server is greater than the rated health value.
  • the processing of step 103 may be as follows: based on the target health value calculation formula and the latest The visit volume information calculates the target health value of the target cache server, and judges whether the target health value is greater than the rated health value; if so, calculate the bandwidth amount of each client to which the latest visit volume information belongs, and follow the order of bandwidth from smallest to largest Select customers as the target to call out clients one by one, and calculate the health value of the target cache server; when the health value of the target cache server is less than or equal to the rated health value, call out clients according to the target for bandwidth scheduling.
  • the background server after the background server obtains the latest traffic information of the target cache server at the beginning of the cycle under each business type, it can calculate the target based on the latest traffic information under each business type and the above-mentioned target health value calculation formula.
  • the health value of the cache server at the beginning of the period may be called the target health value).
  • the background server can determine whether the aforementioned target health value is greater than the rated health value. If the aforementioned target health value is greater than the rated health value, the background server can calculate the bandwidth of each client to which the latest access information belongs, and will sort all the clients in ascending order of bandwidth.
  • the back-end server can select the above-mentioned clients as call-out clients (which can be called target call-out clients) one by one in the order of the amount of bandwidth from small to large, and calculate the health value of the target cache server after each mark, and set the health value Compare with the rated health value.
  • the backend server can stop marking and call out the client to perform bandwidth scheduling on the target cache server according to the above target.
  • the background server can transfer bandwidth to the target cache server when the target health value of the target cache server is less than the rated health value
  • the corresponding processing can be as follows: If not, determine the current time except the target cache server
  • the outgoing client of the cache server selects the outgoing client as the target incoming client in the order of the bandwidth of the outgoing client, and calculates the health value of the target cache server; when the health value of the target cache server is less than or equal to the rated value
  • the client is called to perform bandwidth scheduling according to the target.
  • the calculation and comparison processing of the health value takes up more network resources.
  • the technicians of the CDN service provider can set a threshold (which can be called the preset difference) for the difference between the health value of each cache server and the rated health value.
  • the background server no longer performs the foregoing calculation and comparison processing when the difference between the health value of the cache server and the rated health value is less than or equal to the preset difference.
  • the background server calculates the target health value of the target cache server at the beginning of the period, it can determine whether the aforementioned target health value is greater than the rated health value. If the aforementioned target health value is less than the rated health value, the background server can determine the call-out clients of other cache servers except the target cache server at the current moment, and sort all the call-out clients in descending order of bandwidth. Then, the back-end server can select the above-mentioned call-out clients marked as call-in clients (which can be called target call-in clients) according to the order of bandwidth from small to large, calculate the health value of the target cache server after each mark, and then The health value is compared with the rated health value.
  • call-in clients which can be called target call-in clients
  • the background server When the health value of the target cache server calculated after a certain mark is less than or equal to the rated health value, and the difference between the rated health value and the health value of the target cache server is less than or equal to the preset difference, the background server then The client can be transferred to the target cache server for bandwidth scheduling according to the above goals.
  • the backend server can also mark client C 8 as The target transfer client, that is, both client C 6 and client C 8 are marked as target transfer client, and the health value of the target cache server is calculated.
  • the back-end server can transfer the client (ie client C 6 and client C 8 ) perform bandwidth scheduling on the target cache server.
  • the background server may perform bandwidth transfer when the target health value is less than the warning health value, and the corresponding processing may be as follows: if not, determine whether the target health value is less than the warning health value; when the target health value is less than the warning health value When the health value is used, the outgoing clients of other cache servers except the target cache server are determined at the current moment.
  • the technical personnel of the CDN service provider can provide each cache server through the back-end server Set an alert health value.
  • the background server can also determine whether the target health value is less than the warning health value.
  • the back-end server can determine the call-out clients of the cache servers other than the target cache server at the current moment, and select the call-out clients as the target in descending order of the bandwidth of the call-out clients.
  • the client Call in the client and calculate the health value of the target cache server.
  • the health value of the target cache server is less than or equal to the rated health value, and the difference between the rated health value and the health value of the target cache server is less than or equal to the preset difference value , According to the target transfer to the customer for bandwidth scheduling. On the contrary, the background server does not need to perform any processing.
  • the background server may call the bandwidth of the target cache server according to the business type adjustment gradient corresponding to the target service category of the target cache server.
  • the processing of step 103 is described above. It can be as follows: Calculate the target health value of the target cache server based on the target health value calculation formula and the latest traffic information, and determine whether the target health value is greater than the rated health value; if so, follow the target corresponding to the target business category to which the target cache server belongs The adjustment gradient of the business type successively reduces the traffic information of the target business type, and calculates the health value of the target cache server; when the health value of the target cache server is less than or equal to the rated health value, the reduced traffic information of the target business type is determined Call the target volume, and perform bandwidth scheduling according to the target call volume.
  • the technical personnel of the CDN service provider can divide the business category of each cache server according to the business type through the back-end server, such as a certain cache server mainly
  • the network service provided is a live broadcast service
  • the back-end server can divide the cache server into live broadcast categories and set a corresponding adjustment gradient for each service type, such as reducing/increasing the number of connections each time.
  • the background server after the background server obtains the latest traffic information of each service type of the target cache server at the beginning of the cycle of the preset detection period, it can calculate based on the latest traffic information of each service type and the aforementioned target health value The formula calculates the target health value of the target cache server at the beginning of the period, and determines whether the target health value is greater than the above-mentioned rated health value. If so, the background server can determine the business category to which the target cache server belongs (may be referred to as the target business category), and the adjustment gradient of the business type corresponding to the target business category (may be called the target business type).
  • the background server may successively reduce the traffic information of the target service type according to the adjustment gradient of the target service type, and calculate the health value of the target cache server.
  • the back-end server can determine the aforementioned reduced access volume information of the target business type as the access volume to be called from the target cache server (can be Called the target call-out volume), and perform bandwidth scheduling on the target cache server according to the target call-out volume.
  • the rated health value of the target cache server is 80 as an example.
  • the backend server can obtain All live broadcast connections in the latest traffic information of the target cache server are sorted according to the live broadcast bandwidth corresponding to each live broadcast connection, as shown in Table 5. Then, the backend server can first select the 3 live broadcast connections with the smallest live bandwidth (that is, connection 3, connection 5, and connection 1) and their corresponding live bandwidth and mark the amount of call-out, and calculate the health value of the target cache server. When the calculated health value is still greater than 80, the back-end server can select the 3 live broadcast connections with the smallest live bandwidth (that is, connection 8, connection 2, connection 4) and their The corresponding live broadcast bandwidth is marked as the call-out amount, and the health value of the target cache server is calculated.
  • the back-end server can repeat the above selection of the live broadcast connection and its corresponding live broadcast bandwidth and mark it as a call-out The process of measuring and calculating the health value. On the contrary, if the calculated health value at this time is less than or equal to 80, the back-end server can transfer all the live connections (that is, connection 3, connection 5, connection 1, Connection 8, connection 2, connection 4) and their corresponding live broadcast bandwidth are determined as the target call-out volume, and bandwidth scheduling is performed on the target cache server according to the target call-out volume.
  • the background server can transfer bandwidth to the target cache server according to the business type adjustment gradient corresponding to the target business category of the target cache server.
  • the corresponding processing may be as follows: if not , Determine the current time the call volume of the target service type of the cache server except the target cache server, and based on the call volume according to the adjustment gradient of the target service type, increase the traffic information of the target service type successively, and calculate the target cache server Health value; when the health value of the target cache server is less than or equal to the rated health value, and the difference between the rated health value and the health value of the target cache server is less than or equal to the preset difference, the increased traffic information is determined as the target adjustment Incoming volume, and according to the target transfer volume for bandwidth scheduling.
  • the backend server can determine the rated health value of the target cache server after calculating the target health value of the target cache server at the beginning of the preset detection period, and determine whether the aforementioned target health value is greater than the rated health value . If not, the background server can determine the amount of call out of other cache servers except the target cache server at the current moment, and sort the amount of call out of the aforementioned other cache servers according to the amount of bandwidth from small to large. Then, the background server may gradually increase the traffic information of the target service type according to the adjustment gradient of the target service type, and calculate the health value of the target cache server.
  • the back-end server can set the aforementioned increased target
  • the traffic volume information of the service type is determined as the traffic volume to be transferred to the target cache server (may be referred to as target transfer volume), and bandwidth scheduling is performed on the target cache server according to the target transfer volume.
  • target transfer volume the traffic volume to be transferred to the target cache server
  • bandwidth scheduling is performed on the target cache server according to the target transfer volume.
  • the target cache server belongs to the live broadcast category, and the adjustment gradient corresponding to the live service type is 3 connections, and the rated health value of the target cache server is 80 as an example, suppose the target health value of the target cache server is 50, and the preset difference is 5.
  • the back-end server can determine the call-out volume of all other cache servers except the target cache server at the current moment, and perform the call-out volume of all other cache servers in the descending order of the live bandwidth corresponding to each live broadcast connection For sorting, see Table 6. Then, the background server can first select the 3 live broadcast connections with the smallest live broadcast bandwidth (ie connection 7, connection 2, connection 5) and their corresponding live broadcast bandwidth and mark as the transfer amount, and calculate the health value of the target cache server.
  • the background server can then select the live broadcast bandwidth from all live connections that are not marked as transfer
  • the smallest 3 live connections that is, connection 11, connection 1, connection 9) and their corresponding live bandwidth are marked as the transfer amount, and the health value of the target cache server is calculated. If the health value calculated at this time is still less than the rated health Value, and the difference between the rated health value and the health value is greater than the preset difference value, the background server can repeat the above process of selecting the live connection and its corresponding live bandwidth as the transfer amount and calculating the health value.
  • the back-end server can transfer all live connections that have been marked as transferred (that is, connection 7 , Connection 2, connection 5, connection 11, connection 1, connection 9) and their corresponding live broadcast bandwidth are determined as the target transfer volume, and bandwidth scheduling is performed on the target cache server according to the target transfer volume.
  • Connection 7 Connection 2 Connection 5 Connection 11 Connection 1 Connection 9 Connection 4 Connection 3 Connection 6
  • Connection 8 150 180 195 210 220 240 255 270 290 305 330
  • the background server may call the bandwidth of the target cache server according to the access information corresponding to the target cache server and the health value adjustment range reference table when the target health value is greater than the rated health value.
  • the processing of 103 can be as follows: calculate the target health value of the target cache server based on the target health value calculation formula and the latest visit volume information, and determine whether the target health value is greater than the rated health value; if so, compare the visit volume information of the target cache server with The adjustment range of the health value refers to the corresponding target call-out range selected from the table, the target call-out range is determined as the target call-out volume, and bandwidth scheduling is performed according to the target call-out volume.
  • the technical personnel of the CDN service provider can A reference table for the adjustment range of the corresponding visit volume information and health value is set for each cache server through the background server.
  • the background server After the background server obtains the latest traffic information of the target cache server under each service type at the beginning of the cycle, it can calculate the target cache server based on the latest traffic information under each service type and the above-mentioned target health value calculation formula The target health value at the beginning of the cycle. Then, the background server can determine whether the target health value is greater than the rated health value.
  • the background server can select the corresponding dispatch range from the reference table of the above-mentioned visit volume information and the adjustment range of the health value as the call-out range (which can be called the target call-out range), and determine the target call-out range as the target call-out range The target cache server bandwidth scheduling based on the target call-out volume.
  • the target cache server’s rated health value of 80 refer to Table 7 for the adjustment range of the access information and health value of the target cache server as an example. Assuming the target cache server’s health value is 65, the backend server can determine the target cache The absolute value of the difference between the server's health value and the rated health value is 15. At this time, the background server can select a scheduling range from all the scheduling plans corresponding to the difference range of 10-20.
  • the back-end server can first determine whether there are 3 connections and the corresponding total bandwidth is 450M. Refer to Table 6 for connection 3, connection 4, and connection 6. The sum of the bandwidth of is 450M, therefore, the backend server can determine connection 3, connection 4, connection 6 and the corresponding live bandwidth of these three connections as the target call-out volume, and perform bandwidth scheduling on the target cache server according to the target call-out volume .
  • Connection 1 To Connection 1 Connection 2 Connection 3 Connection 4 Connection 5 Connection 6 Connection 7 Connection 8 Connection 9 Bandwidth (M) 100 150 120 130 160 200 210 190 250
  • the background server may transfer the bandwidth of the target cache server according to the access information and the health value adjustment range reference table corresponding to the target cache server.
  • the corresponding processing may be as follows : If not, select the corresponding target transfer range from the reference table of the access volume information and health value adjustment range of the target cache server, determine the target transfer range as the target transfer volume, and perform bandwidth according to the target transfer volume Scheduling.
  • the background server after the background server calculates the target health value of the target cache server at the beginning of the period, it can determine whether the target health value is greater than the rated health value. If not, the background server can select the corresponding scheduling range from the reference table for the adjustment range of the visit volume information and health value as the transfer range (which can be called the target transfer range), and then the background server can transfer the target The amplitude is determined as the target transfer amount, and bandwidth scheduling can be performed on the target cache server based on the target transfer amount.
  • the preset detection period obtain the latest visit volume information of each service type of the target cache server at the beginning of the period; obtain the historical health value based on the target cache server and historical visits under each service type
  • the target health value calculation formula fitted by the amount information, and the rated health value of the target cache server based on the latest visit volume information, the target health value calculation formula, and the rated health value to the target cache server Perform bandwidth scheduling.
  • the back-end server can determine the health status of the cache server based on the access volume information, health value calculation formula and rated health value of the cache server under each service type, and can perform bandwidth scheduling on the cache server according to the access volume information under each service type , In order to keep the load condition of the cache server in good condition, which can effectively improve the network service quality.
  • an embodiment of the present application also provides a bandwidth scheduling device for a cache server.
  • the device includes an information acquisition module 201 and a scheduling module 202, wherein:
  • the information acquisition module 201 is configured to acquire the latest traffic information of each service type of the target cache server at the beginning of the cycle according to a preset detection cycle;
  • the scheduling module 202 is configured to perform bandwidth scheduling on the target cache server based on the latest traffic information, the target health value calculation formula, and the rated health value.
  • the device further includes a calculation module 203 and a formula update module 204, wherein:
  • the calculation module 203 is configured to obtain statistical data of each performance index of the target cache server in the historical time period before the start of the cycle according to the preset formula update period, and calculate the statistical data based on the statistical data and the preset performance health formula The historical health value of the target cache server at each time within the historical time period;
  • the formula update module 204 is configured to obtain historical visit volume information of each service type of the target cache server in the historical time period before the cycle start time, and based on the historical health value and the historical visit The amount information updates the target health value calculation formula of the target cache server.
  • the formula update module 204 is further used for:
  • the device further includes a judging module for:
  • the scheduling module 202 is also used for:
  • the client When the health value of the target cache server is less than or equal to the rated health value, the client is called for bandwidth scheduling according to the target.
  • the scheduling module 202 is further used for:
  • the health value of the target cache server is less than or equal to the rated health value, and the difference between the rated health value and the health value of the target cache server is less than or equal to a preset difference value, adjust according to the target Incoming customers for bandwidth scheduling.
  • the scheduling module 202 is further used for:
  • the judgment module is used to:
  • the scheduling module 202 is also used for:
  • the reduced access volume information of the target service type is determined as the target call-out volume, and bandwidth scheduling is performed according to the target call-out volume.
  • the scheduling module 202 is further used for:
  • the The increased access volume information is determined as the target transfer volume, and bandwidth scheduling is performed according to the target transfer volume.
  • the judgment module is used to:
  • the scheduling module 202 is also used for:
  • the scheduling module 202 is further used for:
  • the traffic information under each service type includes at least live broadcast bandwidth, live broadcast connections, on-demand bandwidth, on-demand connections, webpage bandwidth, and webpage connections.
  • the preset detection period obtain the latest visit volume information of each service type of the target cache server at the beginning of the period; obtain the historical health value based on the target cache server and historical visits under each service type
  • the target health value calculation formula fitted by the amount information, and the rated health value of the target cache server based on the latest visit volume information, the target health value calculation formula, and the rated health value to the target cache server Perform bandwidth scheduling.
  • the back-end server can determine the health status of the cache server based on the access volume information, health value calculation formula and rated health value of the cache server under each service type, and can perform bandwidth scheduling on the cache server according to the access volume information under each service type , In order to keep the load of the cache server in good condition, which can effectively improve the network service quality.
  • the bandwidth scheduling device of the cache server provided in the above-mentioned embodiment only uses the division of the above-mentioned functional modules as an example.
  • the above-mentioned function allocation can be completed by different functional modules as required, that is, the The internal structure is divided into different functional modules to complete all or part of the functions described above.
  • the bandwidth scheduling device of the cache server provided in the foregoing embodiment and the embodiment of the bandwidth scheduling method of the cache server belong to the same concept. For the specific implementation process, please refer to the method embodiment, which will not be repeated here.
  • Fig. 4 is a schematic structural diagram of a server provided by an embodiment of the present application.
  • the server 400 may have relatively large differences due to different configurations or performance, and may include one or more central processing units 422 (for example, one or more processors) and a memory 432, and one or more storage application programs 442 or data 444 storage medium 430 (for example, one or a storage device with a large amount of storage).
  • the memory 432 and the storage medium 430 may be short-term storage or persistent storage.
  • the program stored in the storage medium 430 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the server 400.
  • the central processing unit 422 may be configured to communicate with the storage medium 430 and execute a series of instruction operations in the storage medium 430 on the server 400.
  • the server 400 may also include one or more power supplies 426, one or more wired or wireless network interfaces 450, one or more input and output interfaces 458, one or more keyboards 456, and/or, one or more operating systems 441 , Such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc.
  • the server 400 may include a memory and one or more programs.
  • One or more programs are stored in the memory and configured to be executed by one or more processors.
  • the bandwidth scheduling instruction of the cache server is not limited to one or more processors.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computer Hardware Design (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Memory System Of A Hierarchy Structure (AREA)

Abstract

本申请提供了一种缓存服务器的带宽调度方法，所述方法包括：按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息(101)；获取基于所述目标缓存服务器的历史健康值及历史访问量信息拟合出的目标健康值计算公式，及所述目标缓存服务器的额定健康值(102)；基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度(103)。采用本申请可以有效提高网络服务质量。

Description

一种缓存服务器的带宽调度方法及装置
交叉引用
本申请引用于2019年07月11日递交的名称为“一种缓存服务器的带宽调度方法及装置”的第201910625761.9号中国专利申请，其通过引用被全部并入本申请。
技术领域
本申请涉及互联网技术领域，特别涉及一种缓存服务器的带宽调度方法及装置。
背景技术
CDN(Content Delivery Network，内容分发网络)服务在现有的互联网中部署多台缓存服务器，使用户可以从一台较近或较优的缓存服务器获取数据，以为用户提供网络加速服务。
为防止由于缓存服务器的实时带宽过高导致缓存服务器的负载过高，进而影响到网络服务质量，CDN服务商可以为部署的每台缓存服务器设置额定带宽，并通过带宽调度保持缓存服务器的实时带宽低于额定带宽。具体来说，CDN服务商的技术人员可以通过后台服务器实时采集各台缓存服务器的实时带宽，并根据每台缓存服务器的实时带宽和额定带宽计算出带宽调出量，即需要调度至其他缓存服务器，由其他缓存服务器提供服务的带宽量，或者计算出带宽调入量，即该缓存服务器还可承载的带宽量，并根据带宽调出/入量进行带宽调度。例如，后台服务器采集到某台缓存服务器的实时带宽为800兆，假设该缓存服务器的额定带宽为700兆，后台服务器可以得出该缓存服务器的带宽调出量为100兆；假设该缓存服务器的额定带宽为1000兆，后台服务器可以得出该缓存 服务器的带宽调入量为200兆。
在实现本申请的过程中，发明人发现现有技术至少存在以下问题：
由于缓存服务器的负载能力不单受承载的带宽总量的影响，还受请求的业务类型、连接数等因素的影响，故而可能会出现按照带宽调出量对某缓存服务器进行带宽调度后，该缓存服务器的负载仍较高，或者按照带宽调入量对某缓存服务器进行带宽调度后，出现该缓存服务器负载过高的现象，进而影响到网络服务质量。
发明内容
为了解决现有技术的问题，本申请实施例提供了一种缓存服务器的带宽调度方法及装置。所述技术方案如下。
第一方面，提供了一种缓存服务器的带宽调度方法，所述方法包括：
按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息；
获取基于所述目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标健康值计算公式，及所述目标缓存服务器的额定健康值；
基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度。
在一个实施例中，所述方法还包括：
按照预设公式更新周期获取目标缓存服务器在周期起始时刻前的历史时长内的各性能指标的统计数据，基于所述统计数据和预设性能健康公式计算所述目标缓存服务器在所述历史时长内各时刻的历史健康值；
获取所述目标缓存服务器在所述周期起始时刻前的历史时长内的各业务类型下的历史访问量信息，并基于所述历史健康值及所述历史访问量信息更新所述目标缓存服务器的目标健康值计算公式。
在一个实施例中，所述基于所述历史健康值及所述历史访问量信息更新 所述目标缓存服务器的目标健康值计算公式，包括：
确定所述目标缓存服务器所属的目标服务器类别，并基于所述历史健康值、所述目标服务器类别及所述历史访问量信息更新所述目标缓存服务器的目标健康值计算公式。
在一个实施例中，所述基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度，包括：
基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
如果是，则计算所述最新访问量信息所属的每个客户的带宽量，并按照所述带宽量从小到大的顺序逐次选取所述客户为目标调出客户，并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值时，按照所述目标调出客户进行带宽调度。
在一个实施例中，所述基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值之后，还包括：
如果否，则确定当前时刻除所述目标缓存服务器外的缓存服务器的调出客户，按照所述调出客户的带宽量从小到大的顺序逐次选取所述调出客户为目标调入客户，并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值，且所述额定健康值与所述目标缓存服务器的健康值的差值小于或等于预设差值时，按照所述目标调入客户进行带宽调度。
在一个实施例中，所述如果否，则确定当前时刻除所述目标缓存服务器外的其他缓存服务器的调出客户，包括：
如果否，则判断所述目标健康值是否小于预警健康值；
当所述目标健康值小于所述预警健康值时，则确定当前时刻除所述目标缓 存服务器外的其他缓存服务器的调出客户。
在一个实施例中，所述基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度，包括：
基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
如果是，则按照所述目标缓存服务器所属的目标业务类别对应的目标业务类型的调整梯度逐次减少所述目标业务类型的访问量信息,并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值时，将所述减少的目标业务类型的访问量信息确定为目标调出量，并按照所述目标调出量进行带宽调度。
在一个实施例中，所述基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值之后，还包括：
如果否，则确定当前时刻除所述目标缓存服务器外的缓存服务器的所述目标业务类型的调出量，并基于所述调出量按照所述目标业务类型的调整梯度逐次增加所述目标业务类型的访问量信息,并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值，且所述额定健康值与所述目标缓存服务器的健康值的差值小于或等于所述预设差值时，将所述增加的访问量信息确定为目标调入量，并按照所述目标调入量进行带宽调度。
在一个实施例中，所述基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度，包括：
基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
如果是，则从所述目标缓存服务器的访问量信息与健康值的调整幅度参照 表中选取对应的目标调出幅度，将所述目标调出幅度确定为目标调出量，并按照所述目标调出量进行带宽调度。
在一个实施例中，所述基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值之后，还包括：
如果否，则从所述目标缓存服务器的访问量信息与健康值的调整幅度参照表中选取对应的目标调入幅度，将所述目标调入幅度确定为目标调入量，并按照所述目标调入量进行带宽调度。
在一个实施例中，所述各业务类型下的访问量信息至少包括直播带宽、直播连接数、点播带宽、点播连接数、网页带宽、网页连接数。
第二方面，提供了一种缓存服务器的带宽调度装置，所述装置包括信息获取模块及调度模块，其中：
所述信息获取模块，用于按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息；
获取基于所述目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标健康值计算公式，及所述目标缓存服务器的额定健康值；
所述调度模块，用于基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度。
在一个实施例中，所述装置还包括计算模块及公式更新模块，其中：
所述计算模块，用于按照预设公式更新周期获取目标缓存服务器在周期起始时刻前的历史时长内的各性能指标的统计数据，基于所述统计数据和预设性能健康公式计算所述目标缓存服务器在所述历史时长内各时刻的历史健康值；
所述公式更新模块，用于获取所述目标缓存服务器在所述周期起始时刻前的历史时长内的各业务类型下的历史访问量信息，并基于所述历史健康值及所述历史访问量信息更新所述目标缓存服务器的目标健康值计算公式。
在一个实施例中，所述公式更新模块，还用于：
确定所述目标缓存服务器所属的目标服务器类别，并基于所述历史健康值、所述目标服务器类别及所述历史访问量信息更新所述目标缓存服务器的目标健康值计算公式。
在一个实施例中，所述装置还包括判断模块，用于：
基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
所述调度模块，还用于：
如果是，则计算所述最新访问量信息所属的每个客户的带宽量，并按照所述带宽量从小到大的顺序逐次选取所述客户为目标调出客户，并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值时，按照所述目标调出客户进行带宽调度。
在一个实施例中，所述调度模块，还用于：
如果否，则确定当前时刻除所述目标缓存服务器外的缓存服务器的调出客户，按照所述调出客户的带宽量从小到大的顺序逐次选取所述调出客户为目标调入客户，并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值，且所述额定健康值与所述目标缓存服务器的健康值的差值小于或等于预设差值时，按照所述目标调入客户进行带宽调度。
在一个实施例中，所述调度模块，还用于：
如果否，则判断所述目标健康值是否小于预警健康值；
当所述目标健康值小于所述预警健康值时，则确定当前时刻除所述目标缓存服务器外的其他缓存服务器的调出客户。
在一个实施例中，所述判断模块，用于：
基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
所述调度模块，还用于：
如果是，则按照所述目标缓存服务器所属的目标业务类别对应的目标业务类型的调整梯度逐次减少所述目标业务类型的访问量信息,并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值时，将所述减少的目标业务类型的访问量信息确定为目标调出量，并按照所述目标调出量进行带宽调度。
在一个实施例中，所述调度模块，还用于：
如果否，则确定当前时刻除所述目标缓存服务器外的缓存服务器的所述目标业务类型的调出量，并基于所述调出量按照所述目标业务类型的调整梯度逐次增加所述目标业务类型的访问量信息,并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值，且所述额定健康值与所述目标缓存服务器的健康值的差值小于或等于所述预设差值时，将所述增加的访问量信息确定为目标调入量，并按照所述目标调入量进行带宽调度。
在一个实施例中，所述判断模块，用于：
基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
所述调度模块，还用于：
如果是，则从所述目标缓存服务器的访问量信息与健康值的调整幅度参照表中选取对应的目标调出幅度，将所述目标调出幅度确定为目标调出量，并按照所述目标调出量进行带宽调度。
在一个实施例中，所述调度模块，还用于：
如果否，则从所述目标缓存服务器的访问量信息与健康值的调整幅度参照表中选取对应的目标调入幅度，将所述目标调入幅度确定为目标调入量，并按照所述目标调入量进行带宽调度。
在一个实施例中，所述各业务类型下的访问量信息至少包括直播带宽、直播连接数、点播带宽、点播连接数、网页带宽、网页连接数。
第三方面，提供了一种服务器，所述服务器包括处理器和存储器，所述存储器中存储有至少一条指令、至少一段程序、代码集或指令集，所述至少一条指令、所述至少一段程序、所述代码集或指令集由所述处理器加载并执行以实现如第一方面所述的缓存服务器的带宽调度方法。
第四方面，提供了一种计算机可读存储介质，所述存储介质中存储有至少一条指令、至少一段程序、代码集或指令集，所述至少一条指令、所述至少一段程序、所述代码集或指令集由处理器加载并执行以实现如第一方面所述的缓存服务器的带宽调度方法。
本申请实施例提供的技术方案带来的有益效果是：在本实施例中，按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息；获取基于所述目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标健康值计算公式，及所述目标缓存服务器的额定健康值；基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度。这样，后台服务器可以基于缓存服务器在各业务类型下的访问量信息、健康值计算公式及额定健康值确定缓存服务器的健康状况，并可以按照各业务类型下的访问量信息对缓存服务器进行带宽调度，以保持缓存服务器的负载状况良好，进而可以有效提高网络服务质量。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其它的附图。
图1是本申请实施例提供的一种缓存服务器的带宽调度方法的流程图；
图2是本申请实施例提供的一种缓存服务器的带宽调度装置的结构示意图；
图3是本申请实施例提供的一种缓存服务器的带宽调度装置的结构示意图；
图4是本申请实施例提供的一种服务器的结构示意图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚，下面将结合附图对本申请实施方式作进一步地详细描述。
本申请实施例提供了一种缓存服务器的带宽调度方法，该方法的执行主体可以是后台服务器，其可以是任意一台服务器，也可以是由多台服务器组成的服务器集群，该后台服务器可以部署在CDN服务商的机房。后台服务器可以根据缓存服务器的健康值计算公式、缓存服务器的额定健康值以及缓存服务器的最新访问量信息，按照各业务类型下的访问量信息对缓存服务器进行带宽调度。上述后台服务器中可以包括处理器、存储器、收发器，处理器可以用于进行下述流程中的对于缓存服务器的带宽调度处理，存储器可以用于存储下述处理过程中需要的数据以及产生的数据，收发器可以用于接收和发送下述处理过程中的相关数据。
下面将结合具体实施方式，对图1所示的一种缓存服务器的带宽调度方法的处理流程进行详细的说明，内容可以如下。
步骤101：按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息。
在实施中，CDN服务商可以通过部署在各个区域的缓存服务器为客户提供网络加速服务，各缓存服务器在为客户提供网络加速服务时，可以记录自身的性能指标(如CPU利用率、内存利用率、连接响应时长、硬盘利用率等)的统计数据，及产生的各业务类型下的访问量信息，其中，各业务类型下的访问 量信息至少可以包括直播带宽、直播连接数、点播带宽、点播连接数、网页带宽、网页连接数等。考虑到不同业务类型的带宽量及连接数可能会对缓存服务器的负载状况产生不同程度的影响，当缓存服务器负载过高时会影响到缓存服务器的网络服务质量，故而，CDN服务商的技术人员可以通过后台服务器以各业务类型及各业务类型对应的连接数为粒度对缓存服务器进行带宽调度。具体的，后台服务器可以按照预设的检测周期获取某台缓存服务器(可称为目标缓存服务器)在当前周期起始时刻已经产生的各业务类型下的最新访问量信息，如在当前周期起始时刻已经产生的直播带宽、直播连接数、点播带宽、点播连接数、网页带宽、网页连接数等最新访问量信息。以使后台服务器可以基于该最新访问量信息、目标缓存服务器的健康值计算公式(可称为目标健康值计算公式)及额定健康值按照各业务类型下的访问量信息对目标缓存服务器进行带宽调度。
步骤102：获取基于目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标健康值计算公式，及目标缓存服务器的额定健康值。
在实施中，后台服务器可以基于每台缓存服务器的历史健康值及每台缓存服务器在各业务类型下的历史访问量信息，为每台缓存服务器拟合一个健康值计算公式，并为每台缓存服务器设置一个额定健康值，以使后台服务器可以通过健康值计算公式计算缓存服务器的健康值，并根据该健康值和额定健康值来确定缓存服务器的负载状况，当缓存服务器的健康值大于额定健康值时，则判定缓存服务器不健康(即负载过高)，反之，则判定缓存服务器健康。后台服务器在获取到目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息之后，可以获取基于目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标缓存服务器的目标健康值计算公式，以及目标缓存服务器的额定健康值。
在一个实施例中，目标缓存服务器可以按照预设公式更新周期更新目标健康值计算公式，相应的处理可以如下：按照预设公式更新周期获取目标缓存 服务器在周期起始时刻前的历史时长内的各性能指标的统计数据，基于统计数据和预设性能健康公式计算目标缓存服务器在历史时长内各时刻的历史健康值。获取目标缓存服务器在周期起始时刻前的历史时长内的各业务类型下的历史访问量信息，并基于历史健康值及历史访问量信息更新目标缓存服务器的目标健康值计算公式。
其中，各业务类型下的历史访问量信息至少包括直播带宽、直播连接数、点播带宽、点播连接数、网页带宽、网页连接数。
在实施中，CDN服务商的技术人员可以在后台服务器为每台缓存服务器预设一个以上述性能指标(如CPU利用率、内存利用率、连接响应时长、硬盘利用率等)为变量的健康值计算公式(可称为预设性能健康公式)，然后，后台服务器可以按照预设的公式更新周期获取目标缓存服务器在当前公式更新周期的周期起始时刻前的历史时长(如可以是1小时)内的各性能指标的统计数据，并基于该统计数据和目标缓存服务器的预设性能健康公式计算出目标缓存服务器在上述历史时长内各时刻的历史健康值(如可以以1分钟为粒度计算目标缓存服务器的历史健康值)。以目标缓存服务器的预设性能健康计算公式为QOS＝ax+by+cz+dk为例，其中QOS表示健康值，x表示CPU利用率，y表示内存利用率，z表示连接响应时长，k表示硬盘利用率，a、b、c、d分别表示每个性能指标对应的系数，假设该缓存服务器在10:00的CPU利用率、内存利用率、连接响应时长、硬盘利用率分别为70％、75％、0.1s、80％，a、b、c、d分别为40、30、20、30则可以得出该缓存服务器在10:00的健康值为HV＝0.7×40+0.75×30+0.1×20+0.8×30＝76.5。
后台服务器在计算出目标缓存服务器在上述历史时长内各时刻的历史健康值之后，可以获取目标缓存服务器在上述周期起始时刻前的历史时长内的各业务类型下的历史访问量信息，并基于该历史访问量信息统计出与上述各历史健康值对应的各时刻的历史访问量信息。然后，后台服务器可以通过机器学习中的回归分析模型(如线性回归、多项式回归等)基于前述统计出的各时刻对 应的各业务类型下的历史访问量信息及上述各时刻对应的历史健康值拟合出目标缓存服务器的健康值计算公式，并将该健康值计算公式更新为目标缓存服务器的目标健康值计算公式。
在一个实施例中，后台服务器还可以结合目标缓存服务器的服务器类别更新目标健康值计算公式，相应的处理可以如下：确定目标缓存服务器所属的目标服务器类别，并基于历史健康值、目标服务器类别及历史访问量信息更新目标缓存服务器的目标健康值计算公式。
在实施中，考虑到缓存服务器的不同硬件配置(如CPU、内存、硬盘等)可能会对缓存服务器的负载能力产生不同程度的影响，故而，CDN服务商的技术人员可以通过后台服务器按照缓存服务器的硬件配置将缓存服务器划分为不同的服务器类别，如可以将CPU、内存、硬盘等硬件配置均相同的缓存服务器划分以同一个服务器类别。这样，后台服务器在获取到上述目标缓存服务器在公式更新周期的周期起始时刻前的历史时长内的各业务类型下的历史访问量信息之后，可以确定目标缓存服务器所属的服务器类别(可称为目标服务器类别)，以根据该目标服务器类别、上述目标缓存服务器在历史时长内各时刻的历史健康值以及目标缓存服务器在上述周期起始时刻前的历史时长内的各业务类型下的历史访问量信息，更新目标缓存服务器的目标健康值计算公式。
步骤103：基于最新访问量信息、目标健康值计算公式及额定健康值对目标缓存服务器进行带宽调度。
在实施中，后台服务器在获取到目标缓存服务器的最新访问量信息、目标健康值计算公式及额定健康值之后，可以基于前述目标缓存服务器的最新访问量信息、前述目标缓存服务器的目标健康值计算公式计算目标缓存服务器的健康值。然后，后台服务器可以根据该目标缓存服务器的健康值、前述目标缓存服务器的额定健康值、目标健康值计算公式以及上述最新访问量信息对目标缓存服务器进行带宽调度。以目标缓存服务器的目标健康值计算公式为：QOS＝a 1x 1+a 2x 2+a 3x 3+b 1x 4+b 2x 5+b 3x 6，其中，QOS表示目标缓存服务器的健康值，x 1 表示目标缓存服务器当前时刻的直播带宽、x 2表示目标缓存服务器当前时刻的点播带宽、x 3表示目标缓存服务器当前时刻的网页带宽、x 4表示目标缓存服务器当前时刻的直播连接数、x 5表示目标缓存服务器当前时刻的点播连接数、x 6表示目标缓存服务器当前时刻的网页连接数，a 1、a 2、a 3、b 1、b 2、b 3分别为各业务类型下的访问量信息对应的系数，假设目标缓存服务器在10:00的各业务类型下的最新访问量信息分别为直播带宽800兆、点播带宽700兆、网页带宽900兆，直播连接数20、点播连接数15、网页连接数40，a 1、a 2、a 3、b 1、b 2、b 3分别为0.03、0.04、0.02、0.35、0.2、0.1，后台服务器则可以基于上述目标健康值计算公式计算出目标缓存服务器的健康值为QOS＝0.03×800+0.04×700+0.02×900+0.35×20+0.2×15+0.1×40＝84。
在一个实施例中，后台服务器可以在目标缓存服务器的目标健康值大于额定健康值时对目标缓存服务器进行带宽调出，相应的，上述步骤103的处理可以如下：基于目标健康值计算公式及最新访问量信息计算目标缓存服务器的目标健康值，并判断目标健康值是否大于额定健康值；如果是，则计算最新访问量信息所属的每个客户的带宽量，并按照带宽量从小到大的顺序逐次选取客户为目标调出客户，并计算目标缓存服务器的健康值；当目标缓存服务器的健康值小于或等于额定健康值时，按照目标调出客户进行带宽调度。
在实施中，后台服务器在获取到目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息之后，可以基于前述各业务类型下的最新访问量信息及上述目标健康值计算公式计算目标缓存服务器在该周期起始时刻的健康值(可称为目标健康值)。然后，后台服务器可以判断前述目标健康值是否大于该额定健康值。如果前述目标健康值大于额定健康值，后台服务器则可以计算上述最新访问量信息所属的每个客户的带宽量，并将按照带宽量从小到大的顺序对上述所有客户进行排序。然后，后台服务器可以按照带宽量从小到大的顺序逐次选取上述客户标记为调出客户(可称为目标调出客户)，在每次标记后计算目标缓存服务器的健康值，并将该健康值与额定健康值进行比较。当在某次标 记目标调出客户后计算出的目标缓存服务器的健康值小于或等于额定健康值时，后台服务器则可以停止标记，并按照上述目标调出客户对目标缓存服务器进行带宽调度。仍以目标缓存服务器的目标健康值计算公式为QOS＝0.03x 1+0.04x 2+0.02x 3+0.35x 4+0.2x 5+0.1x 6为例，假设目标缓存服务器的额定健康值为75，最新访问量信息所属的客户有C 1、C 2、C 3，且客户C 1、C 2、C 3对应的各业务类型下的访问量信息见表1。
表1
Figure PCTCN2020074968-appb-000001
后台服务器可以计算出目标缓存服务器的目标健康值为，QOS＝0.03×(250+400+280)+0.04×(180+500+320)+0.02×(170+400+380)+0.35×(5+8+7)+0.2×(7+10+8)+0.1×(9+12+15)＝102.5，可见，目标缓存服务器的目标健康值远大于额定健康值。然后，后台服务器可以计算出上述最新访问量信息所属的每个客户的带宽量，并按照带宽量从小到大的顺序进行排序，见表2。
表2
Figure PCTCN2020074968-appb-000002
然后，后台服务器可以先将客户C 1标记为目标调出客户，并计算目标缓存服务器的健康值，此时QOS＝0.03×(280+400)+0.04×(320+500)+0.02×(380+400)+0.35×(7+8)+0.2×(8+10)+0.1×(15+12)＝80.35。由于该健康值仍大于额定健康值，后台服务器可以将客户C 3也标记为目标调出客户，即客户C 1和客户C 3均被标记为目标调出客户，并计算目标缓存服务器的健康值，此 时，QOS＝0.03×400+0.04×500+0.02×400+0.35×8+0.2×10+0.1×12＝46，由于该健康值小于额定健康值，所以，后台服务器可以按照上述目标调出客户(即客户C 1和客户C 3)对目标缓存服务器进行带宽调度。
在一个实施例中，后台服务器可以在目标缓存服务器的目标健康值小于额定健康值时对目标缓存服务器进行带宽调入，相应的处理可以如下：如果否，则确定当前时刻除目标缓存服务器外的缓存服务器的调出客户，按照调出客户的带宽量从小到大的顺序逐次选取调出客户为目标调入客户，并计算目标缓存服务器的健康值；当目标缓存服务器的健康值小于或等于额定健康值，且额定健康值与目标缓存服务器的健康值的差值小于或等于预设差值时，按照目标调入客户进行带宽调度。
在实施中，考虑到在进行带宽调度时，当缓存服务器的健康值与额定健康值的差值较小时，可能不能再承载其他客户的访问量，故而，为避免后台服务器在这种情况下进行健康值的计算和比较处理占用较多的网络资源，CDN服务商的技术人员可以为每台缓存服务器的健康值与额定健康值的差值设置一个阈值(可称为预设差值)，以使后台服务器在缓存服务器的健康值与额定健康值的差值小于或等于预设差值时，则不再进行上述计算和比较处理。具体的，后台服务器计算出目标缓存服务器在该周期起始时刻的目标健康值之后，可以判断前述目标健康值是否大于额定健康值。如果前述目标健康值小于额定健康值，后台服务器则可以确定当前时刻除目标缓存服务器外的其他缓存服务器的调出客户，并按照带宽量从小到大的顺序对上述所有调出客户进行排序。然后，后台服务器可以按照带宽量从小到大的顺序依次选取上述调出客户标记为调入客户(可称为目标调入客户)，在每次标记后计算目标缓存服务器的健康值，并将该健康值与额定健康值进行比较。当在某次标记后计算出的目标缓存服务器的健康值小于或等于额定健康值，且额定健康值与所述目标缓存服务器的健康值的差值小于或等于预设差值时，后台服务器则可以按照上述目标调入客户对目标缓存服务器进行带宽调度。仍以目标缓存服务器的目标健康值计算公式为 QOS＝0.03x 1+0.04x 2+0.02x 3+0.35x 4+0.2x 5+0.1x 6为例，假设目标缓存服务器的额定健康值为85，最新访问量信息所属的客户有C 4、C 5，且客户C 4、C 5对应的各业务类型下的访问量信息见表3。
表3
Figure PCTCN2020074968-appb-000003
后台服务器可以计算出目标缓存服务器的目标健康值为，QOS＝0.03×(200+300)+0.04×(150+280)+0.02×(210+320)+0.35×(5+9)+0.2×(6+7)+0.1×(9+10)＝52.2，可见，目标缓存服务器的目标健康值小于额定健康值。然后，后台服务器可以确定当前时刻除目标缓存服务器外的其他缓存服务器的调出客户，假设其他缓存服务器的调出客户为C 6、C 7、C 8，按照带宽量从小到大的顺序对C 6、C 7、C 8进行排序，见表4。
表4
Figure PCTCN2020074968-appb-000004
然后，后台服务器可以先将客户C 6标记为目标调入客户，并计算目标缓存服务器的健康值，此时QOS＝0.03×(200+300+150)+0.04×(150+280+120)+0.02×(210+320+130)+0.35×(5+9+3)+0.2×(6+7+4)+0.1×(9+10+3)＝66.25。假设目标缓存服务器的预设差值为5，由于该健康值仍小于额定健康值，且额定健康值与该健康值的差值大于预设差值，故而后台服务器可以将客户C 8也标记为目标调入客户，即客户C 6和客户C 8均被标记为目标调入客户，并计算目标缓存服务器的健康值，此时，QOS＝0.03×(200+300+150+190)+0.04×(150+280+120+180)+0.02×(210+320+130+130)+0.35×(5+9+3+4)+0.2× (6+7+4+3)+0.1×(9+10+3+3)＝84.05，由于该健康值小于额定健康值，且额定健康值与该健康值的差值小于预设差值，所以后台服务器可以按照上述目标调入客户(即客户C 6和客户C 8)对目标缓存服务器进行带宽调度。
在一个实施例中，后台服务器可以在上述目标健康值小于预警健康值时进行带宽调入，相应的处理可以如下：如果否，则判断目标健康值是否小于预警健康值；当目标健康值小于预警健康值时，则确定当前时刻除目标缓存服务器外的其他缓存服务器的调出客户。
在实施中，考虑到在上述目标健康值小于额定健康值时，后台服务器进行的上述处理过程可能会占用较多的网络资源，故而，CDN服务商的技术人员可以通过后台服务器为每台缓存服务器设置一个预警健康值。这样，当上述目标缓存服务器的目标健康值小于额定健康值时，后台服务器还可以判断该目标健康值是否小于预警健康值。当该目标健康值小于预警健康值时，后台服务器则可以确定当前时刻除目标缓存服务器外的缓存服务器的调出客户，按照调出客户的带宽量从小到大的顺序逐次选取调出客户为目标调入客户，并计算目标缓存服务器的健康值，在目标缓存服务器的健康值小于或等于额定健康值，且所额定健康值与目标缓存服务器的健康值的差值小于或等于预设差值时，按照目标调入客户进行带宽调度。反之，后台服务器则无需进行任何处理。
在一个实施例中，后台服务器可以在目标健康值大于额定健康值时，按照目标缓存服务器的目标业务类别对应的业务类型调整梯度对目标缓存服务器进行带宽调出，相应的，上述步骤103的处理可以如下：基于目标健康值计算公式及最新访问量信息计算目标缓存服务器的目标健康值，并判断目标健康值是否大于额定健康值；如果是，则按照目标缓存服务器所属的目标业务类别对应的目标业务类型的调整梯度逐次减少目标业务类型的访问量信息,并计算目标缓存服务器的健康值；当目标缓存服务器的健康值小于或等于额定健康值时，将减少的目标业务类型的访问量信息确定为目标调出量，并按照目标调出量进行带宽调度。
在实施中，由于每台缓存服务器主要提供的网络服务的业务类型可能不同，故而，CDN服务商的技术人员可以通过后台服务器按照业务类型对每台缓存服务器划分业务类别，如某台缓存服务器主要提供的网络服务为直播业务，后台服务器则可以将该缓存服务器划分为直播类别，并为每种业务类型设置对应的调整梯度，如可以每次减少/增加几条连接数。这样，后台服务器在获取到目标缓存服务器在预设检测周期的周期起始时刻的各业务类型下的最新访问量信息之后，可以基于该各业务类型下的最新访问量信息及上述目标健康值计算公式计算目标缓存服务器在该周期起始时刻的目标健康值，并判断该目标健康值是否大于上述额定健康值。如果是，后台服务器则可以确定目标缓存服务器所属的业务类别(可称为目标业务类别)，及该目标业务类别对应的业务类型(可称为目标业务类型)的调整梯度。然后，后台服务器可以按照该目标业务类型的调整梯度逐次减少上述目标业务类型的访问量信息，并计算目标缓存服务器的健康值。当后台服务器计算出的目标缓存服务器的健康值小于或等于上述额定健康值时，后台服务器则可以将前述减少的目标业务类型的访问量信息确定为要从目标缓存服务器调出的访问量(可称为目标调出量)，并按照目标调出量对该目标缓存服务器进行带宽调度。以目标缓存服务器属于直播类别，且直播业务类型对应的调整梯度为3条连接，目标缓存服务器的额定健康值为80为例，假设目标缓存服务器的目标健康值为90，后台服务器则可以将获取到的目标缓存服务器的最新访问量信息中的所有直播连接按照每条直播连接对应的直播带宽从小到大进行排序，见表5。然后，后台服务器可以先选取直播带宽最小的3条直播连接(即连接3、连接5、连接1)及其对应的直播带宽及标记为调出量，并计算目标缓存服务器的健康值，如果此时计算出的健康值仍大于80，后台服务器则可以在未被标记为调出量的所有直播连接中再选取直播带宽最小的3条直播连接(即连接8、连接2、连接4)及其对应的直播带宽标记为调出量，并计算目标缓存服务器的健康值，如果此时计算出的健康值仍大于80，后台服务器则可以重复上述选取直播连接及其对应的直播带宽标记为调出量并计算健康 值的过程，反之，如果此时计算出的健康值小于或等于80，后台服务器则可以将所有已被标记为调出量的直播连接(即连接3、连接5、连接1、连接8、连接2、连接4)及其对应的直播带宽确定为目标调出量，并按照目标调出量对目标缓存服务器进行带宽调度。
表5
连接3 连接5 连接1 连接8 连接2 连接4 连接6 连接11 连接10 连接7 连接9
150 180 195 210 220 240 255 270 290 305 330
在一个实施例中，后台服务器可以在目标健康值小于额定健康值时，按照目标缓存服务器的目标业务类别对应的业务类型调整梯度对目标缓存服务器进行带宽调入，相应的处理可以如下：如果否，则确定当前时刻除目标缓存服务器外的缓存服务器的目标业务类型的调出量，并基于调出量按照目标业务类型的调整梯度逐次增加目标业务类型的访问量信息,并计算目标缓存服务器的健康值；当目标缓存服务器的健康值小于或等于额定健康值，且额定健康值与目标缓存服务器的健康值的差值小于或等于预设差值时，将增加的访问量信息确定为目标调入量，并按照目标调入量进行带宽调度。
在实施中，后台服务器在计算出目标缓存服务器在预设检测周期的周期起始时刻的目标健康值之后，可以确定该目标缓存服务器的额定健康值，并判断前述目标健康值是否大于额定健康值。如果否，后台服务器则可以确定当前时刻除目标缓存服务器外的其他缓存服务器的调出量，并按照带宽量对前述其他缓存服务器的调出量从小到大进行排序。然后，后台服务器可以按照前述目标业务类型的调整梯度，逐次增加上述目标业务类型的访问量信息，并计算目标缓存服务器的健康值。当后台服务器计算出的目标缓存服务器的健康值小于或等于上述额定健康值，且额定健康值与该健康值的差值小于或等于上述预设差值时，后台服务器则可以将前述增加的目标业务类型的访问量信息确定为要调入该目标缓存服务器的访问量(可称为目标调入量)，并按照目标调入量对该目标缓存服务器进行带宽调度。仍以目标缓存服务器属于直播类别，且直播业务类型对应的调整梯度为3条连接，目标缓存服务器的额定健康值为80为例， 假设目标缓存服务器的目标健康值为50，预设差值为5，后台服务器则可以确定当前时刻除目标缓存服务器之外的所有其他缓存服务器的调出量，并将所有其他缓存服务器的调出量按照每条直播连接对应的直播带宽从小到大的顺序进行排序，见表6。然后，后台服务器可以先选取直播带宽最小的3条直播连接(即连接7、连接2、连接5)及其对应的直播带宽及标记为调入量，并计算目标缓存服务器的健康值，如果此时计算出的健康值仍小于额定健康值，且额定健康值与该健康值的差值大于预设差值，后台服务器则可以在未被标记为调入量的所有直播连接中再选取直播带宽最小的3条直播连接(即连接11、连接1、连接9)及其对应的直播带宽标记为调入量，并计算目标缓存服务器的健康值，如果此时计算出的健康值仍小于额定健康值，且额定健康值与该健康值的差值大于预设差值，后台服务器则可以重复上述选取直播连接及其对应的直播带宽标记为调入量并计算健康值的过程，反之，如果此时计算出的健康值小于或等于80，且额定健康值与该健康值的差值小于或等于预设差值，后台服务器则可以将所有已被标记为调入量的直播连接(即连接7、连接2、连接5、连接11、连接1、连接9)及其对应的直播带宽确定为目标调入量，并按照目标调入量对目标缓存服务器进行带宽调度。
表6
连接7 连接2 连接5 连接11 连接1 连接9 连接4 连接3 连接6 连接10 连接8
150 180 195 210 220 240 255 270 290 305 330
在一个实施例中，后台服务器可以在目标健康值大于额定健康值时，按照目标缓存服务器对应的访问量信息与健康值的调整幅度参照表对目标缓存服务器进行带宽调出，相应的，上述步骤103的处理可以如下：基于目标健康值计算公式及最新访问量信息计算目标缓存服务器的目标健康值，并判断目标健康值是否大于额定健康值；如果是，则从目标缓存服务器的访问量信息与健康值的调整幅度参照表中选取对应的目标调出幅度，将目标调出幅度确定为目标调出量，并按照目标调出量进行带宽调度。
在实施中，考虑到上述基于目标健康值计算公式、额定健康值及最新访问量信息对目标缓存服务器进行带宽调度的处理，可能会占用较多的网络资源，故而，CDN服务商的技术人员可以通过后台服务器为每台缓存服务器设置一个对应的访问量信息与健康值的调整幅度参照表。这样，后台服务器在获取到目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息之后，可以基于该各业务类型下的最新访问量信息及上述目标健康值计算公式计算目标缓存服务器在该周期起始时刻的目标健康值。然后，后台服务器可以判断目标健康值是否大于所述额定健康值。如果是，后台服务器则可以从上述访问量信息与健康值的调整幅度参照表中选取对应的调度幅度作为调出幅度(可称为目标调出幅度)，将该目标调出幅度确定为目标调出量，并基于该目标调出量对目标缓存服务器进行带宽调度。以目标缓存服务器的额定健康值为80，目标缓存服务器对应的访问量信息与健康值的调整幅度参照表见表7为例，假设目标缓存服务器的健康值为65，后台服务器则可以确定目标缓存服务器的健康值与额定健康值的差值的绝对值为15，此时，后台服务器可以在差值范围10-20对应的所有调度方案中选取一个调度幅度。
表7
Figure PCTCN2020074968-appb-000005
假设目标缓存服务器的各业务类型下的访问量信息如表8所示，后台服务器则可以优先判断是否存在3条连接其对应的总带宽为450M，参见表6可见连接3、连接4及连接6的带宽之和为450M，故而，后台服务器可以将连接3、 连接4、连接6及这三条连接对应的直播带宽确定为目标调出量，并按照该目标调出量对目标缓存服务器进行带宽调度。
表8
  连接1 连接2 连接3 连接4 连接5 连接6 连接7 连接8 连接9
带宽(M) 100 150 120 130 160 200 210 190 250
在一个实施例中，后台服务器可以在目标健康值小于额定健康值时，按照目标缓存服务器对应的访问量信息与健康值的调整幅度参照表对目标缓存服务器进行带宽调入，相应的处理可以如下：如果否，则从目标缓存服务器的访问量信息与健康值的调整幅度参照表中选取对应的目标调入幅度，将目标调入幅度确定为目标调入量，并按照目标调入量进行带宽调度。
在实施中，后台服务器在计算出目标缓存服务器在该周期起始时刻的目标健康值之后，可以判断目标健康值是否大于所述额定健康值。如果否，后台服务器则可以从上述访问量信息与健康值的调整幅度参照表中选取对应的调度幅度作为调入幅度(可称为目标调入幅度)，然后，后台服务器可以将该目标调入幅度确定为目标调入量，并可以基于该目标调入量对目标缓存服务器进行带宽调度。
在本实施例中，按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息；获取基于所述目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标健康值计算公式，及所述目标缓存服务器的额定健康值；基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度。这样，后台服务器可以基于缓存服务器在各业务类型下的访问量信息、健康值计算公式及额定健康值确定缓存服务器的健康状况，并可以按照各业务类型下的访问量信息对缓存服务器进行带宽调度，以保持缓存服务器的负载状况良好，进而可以有效提高网络服务质量。
基于相同的技术构思，本申请实施例还提供了一种缓存服务器的带宽调 度装置，如图2所示，所述装置包括信息获取模块201及调度模块202，其中：
所述信息获取模块201，用于按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息；
获取基于所述目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标健康值计算公式，及所述目标缓存服务器的额定健康值；
所述调度模块202，用于基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度。
在一个实施例中，如图3所示，所述装置还包括计算模块203及公式更新模块204，其中：
所述计算模块203，用于按照预设公式更新周期获取目标缓存服务器在周期起始时刻前的历史时长内的各性能指标的统计数据，基于所述统计数据和预设性能健康公式计算所述目标缓存服务器在所述历史时长内各时刻的历史健康值；
所述公式更新模块204，用于获取所述目标缓存服务器在所述周期起始时刻前的历史时长内的各业务类型下的历史访问量信息，并基于所述历史健康值及所述历史访问量信息更新所述目标缓存服务器的目标健康值计算公式。
在一个实施例中，所述公式更新模块204，还用于：
确定所述目标缓存服务器所属的目标服务器类别，并基于所述历史健康值、所述目标服务器类别及所述历史访问量信息更新所述目标缓存服务器的目标健康值计算公式。
在一个实施例中，所述装置还包括判断模块，用于：
基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
所述调度模块202，还用于：
如果是，则计算所述最新访问量信息所属的每个客户的带宽量，并按照所述带宽量从小到大的顺序逐次选取所述客户为目标调出客户，并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值时，按照所述目标调出客户进行带宽调度。
在一个实施例中，所述调度模块202，还用于：
如果否，则确定当前时刻除所述目标缓存服务器外的缓存服务器的调出客户，按照所述调出客户的带宽量从小到大的顺序逐次选取所述调出客户为目标调入客户，并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值，且所述额定健康值与所述目标缓存服务器的健康值的差值小于或等于预设差值时，按照所述目标调入客户进行带宽调度。
在一个实施例中，所述调度模块202，还用于：
如果否，则判断所述目标健康值是否小于预警健康值；
当所述目标健康值小于所述预警健康值时，则确定当前时刻除所述目标缓存服务器外的其他缓存服务器的调出客户。
在一个实施例中，所述判断模块，用于：
基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
所述调度模块202，还用于：
如果是，则按照所述目标缓存服务器所属的目标业务类别对应的目标业务类型的调整梯度逐次减少所述目标业务类型的访问量信息,并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值时，将所述减少的目标业务类型的访问量信息确定为目标调出量，并按照所述目标调出量进行带宽调度。
在一个实施例中，所述调度模块202，还用于：
如果否，则确定当前时刻除所述目标缓存服务器外的缓存服务器的所述目标业务类型的调出量，并基于所述调出量按照所述目标业务类型的调整梯度逐 次增加所述目标业务类型的访问量信息,并计算所述目标缓存服务器的健康值；
当所述目标缓存服务器的健康值小于或等于所述额定健康值，且所述额定健康值与所述目标缓存服务器的健康值的差值小于或等于所述预设差值时，将所述增加的访问量信息确定为目标调入量，并按照所述目标调入量进行带宽调度。
在一个实施例中，所述判断模块，用于：
基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
所述调度模块202，还用于：
如果是，则从所述目标缓存服务器的访问量信息与健康值的调整幅度参照表中选取对应的目标调出幅度，将所述目标调出幅度确定为目标调出量，并按照所述目标调出量进行带宽调度。
在一个实施例中，所述调度模块202，还用于：
如果否，则从所述目标缓存服务器的访问量信息与健康值的调整幅度参照表中选取对应的目标调入幅度，将所述目标调入幅度确定为目标调入量，并按照所述目标调入量进行带宽调度。
在一个实施例中，所述各业务类型下的访问量信息至少包括直播带宽、直播连接数、点播带宽、点播连接数、网页带宽、网页连接数。
在本实施例中，按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息；获取基于所述目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标健康值计算公式，及所述目标缓存服务器的额定健康值；基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度。这样，后台服务器可以基于缓存服务器在各业务类型下的访问量信息、健康值计算公式及额定健康值确定缓存服务器的健康状况，并可以按照各业务类型下的访问量信息对缓存服务器进行带宽调度，以保持缓存服务器的负载状况良好，进而可以有效 提高网络服务质量。
需要说明的是：上述实施例提供的缓存服务器的带宽调度装置仅以上述各功能模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能模块完成，即将装置的内部结构划分成不同的功能模块，以完成以上描述的全部或者部分功能。另外，上述实施例提供的缓存服务器的带宽调度装置与缓存服务器的带宽调度方法实施例属于同一构思，其具体实现过程详见方法实施例，这里不再赘述。
图4是本申请实施例提供的服务器的结构示意图。该服务器400可因配置或性能不同而产生比较大的差异，可以包括一个或一个以***处理器422(例如，一个或一个以上处理器)和存储器432，一个或一个以上存储应用程序442或数据444的存储介质430(例如一个或一个以上海量存储设备)。其中，存储器432和存储介质430可以是短暂存储或持久存储。存储在存储介质430的程序可以包括一个或一个以上模块(图示没标出)，每个模块可以包括对服务器400中的一系列指令操作。更进一步地，中央处理器422可以设置为与存储介质430通信，在服务器400上执行存储介质430中的一系列指令操作。
服务器400还可以包括一个或一个以上电源426，一个或一个以上有线或无线网络接口450，一个或一个以上输入输出接口458，一个或一个以上键盘456，和/或，一个或一个以上操作***441，例如Windows Server，Mac OS X，Unix，Linux，FreeBSD等等。
服务器400可以包括有存储器，以及一个或者一个以上的程序，其中一个或者一个以上程序存储于存储器中，且经配置以由一个或者一个以上处理器执行所述一个或者一个以上程序包含用于进行上述缓存服务器的带宽调度指令。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成，也可以通过程序来指令相关的硬件完成，所述的程序可以存储于一种计算机可读存储介质中，上述提到的存储介质可以是只读存储器，磁盘或光盘等。
以上所述仅为本申请的较佳实施例，并不用以限制本申请，凡在本申请的精神和原则之内，所作的任何修改、等同替换、改进等，均应包含在本申请的保护范围之内。

Claims (14)

  1. 一种缓存服务器的带宽调度方法，包括：
    按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息；
    获取基于所述目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标健康值计算公式，及所述目标缓存服务器的额定健康值；
    基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度。
  2. 根据权利要求1所述的方法，其中，所述方法还包括：
    按照预设公式更新周期获取目标缓存服务器在周期起始时刻前的历史时长内的各性能指标的统计数据，基于所述统计数据和预设性能健康公式计算所述目标缓存服务器在所述历史时长内各时刻的历史健康值；
    获取所述目标缓存服务器在所述周期起始时刻前的历史时长内的各业务类型下的历史访问量信息，并基于所述历史健康值及所述历史访问量信息更新所述目标缓存服务器的目标健康值计算公式。
  3. 根据权利要求2所述的方法，其中，所述基于所述历史健康值及所述历史访问量信息更新所述目标缓存服务器的目标健康值计算公式，包括：
    确定所述目标缓存服务器所属的目标服务器类别，并基于所述历史健康值、所述目标服务器类别及所述历史访问量信息更新所述目标缓存服务器的目标健康值计算公式。
  4. 根据权利要求1所述的方法，其中，所述基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度，包括：
    基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
    如果是，则计算所述最新访问量信息所属的每个客户的带宽量，并按照所述带宽量从小到大的顺序逐次选取所述客户为目标调出客户，并计算所述目标缓存服务器的健康值；
    当所述目标缓存服务器的健康值小于或等于所述额定健康值时，按照所述目标调出客户进行带宽调度。
  5. 根据权利要求4所述的方法，其中，所述基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值之后，还包括：
    如果否，则确定当前时刻除所述目标缓存服务器外的缓存服务器的调出客户，按照所述调出客户的带宽量从小到大的顺序逐次选取所述调出客户为目标调入客户，并计算所述目标缓存服务器的健康值；
    当所述目标缓存服务器的健康值小于或等于所述额定健康值，且所述额定健康值与所述目标缓存服务器的健康值的差值小于或等于预设差值时，按照所述目标调入客户进行带宽调度。
  6. 根据权利要求5所述的方法，其中，所述如果否，则确定当前时刻除所述目标缓存服务器外的其他缓存服务器的调出客户，包括：
    如果否，则判断所述目标健康值是否小于预警健康值；
    当所述目标健康值小于所述预警健康值时，则确定当前时刻除所述目标缓存服务器外的其他缓存服务器的调出客户。
  7. 根据权利要求1所述的方法，其中，所述基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度，包括：
    基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
    如果是，则按照所述目标缓存服务器所属的目标业务类别对应的目标业务类型的调整梯度逐次减少所述目标业务类型的访问量信息,并计算所述目标缓 存服务器的健康值；
    当所述目标缓存服务器的健康值小于或等于所述额定健康值时，将所述减少的目标业务类型的访问量信息确定为目标调出量，并按照所述目标调出量进行带宽调度。
  8. 根据权利要求7所述的方法，其中，所述基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值之后，还包括：
    如果否，则确定当前时刻除所述目标缓存服务器外的缓存服务器的所述目标业务类型的调出量，并基于所述调出量按照所述目标业务类型的调整梯度逐次增加所述目标业务类型的访问量信息,并计算所述目标缓存服务器的健康值；
    当所述目标缓存服务器的健康值小于或等于所述额定健康值，且所述额定健康值与所述目标缓存服务器的健康值的差值小于或等于所述预设差值时，将所述增加的访问量信息确定为目标调入量，并按照所述目标调入量进行带宽调度。
  9. 根据权利要求1所述的方法，其中，所述基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度，包括：
    基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值；
    如果是，则从所述目标缓存服务器的访问量信息与健康值的调整幅度参照表中选取对应的目标调出幅度，将所述目标调出幅度确定为目标调出量，并按照所述目标调出量进行带宽调度。
  10. 根据权利要求9所述的方法，其中，所述基于所述目标健康值计算公式及所述最新访问量信息计算所述目标缓存服务器的目标健康值，并判断所述目标健康值是否大于所述额定健康值之后，还包括：
    如果否，则从所述目标缓存服务器的访问量信息与健康值的调整幅度参照 表中选取对应的目标调入幅度，将所述目标调入幅度确定为目标调入量，并按照所述目标调入量进行带宽调度。
  11. 根据权利要求1或2所述的方法，其中，所述各业务类型下的访问量信息至少包括直播带宽、直播连接数、点播带宽、点播连接数、网页带宽、网页连接数。
  12. 一种缓存服务器的带宽调度装置，包括信息获取模块及调度模块，其中：
    所述信息获取模块，用于按照预设检测周期获取目标缓存服务器在周期起始时刻的各业务类型下的最新访问量信息；
    获取基于所述目标缓存服务器的历史健康值及各业务类型下的历史访问量信息拟合出的目标健康值计算公式，及所述目标缓存服务器的额定健康值；
    所述调度模块，用于基于所述最新访问量信息、所述目标健康值计算公式及所述额定健康值对所述目标缓存服务器进行带宽调度。
  13. 一种服务器，包括处理器和存储器，所述存储器中存储有至少一条指令、至少一段程序、代码集或指令集，所述至少一条指令、所述至少一段程序、所述代码集或指令集由所述处理器加载并执行以实现如权利要求1至11任一所述的缓存服务器的带宽调度方法。
  14. 一种计算机可读存储介质，存储有至少一条指令、至少一段程序、代码集或指令集，所述至少一条指令、所述至少一段程序、所述代码集或指令集由处理器加载并执行以实现如权利要求1至11任一所述的缓存服务器的带宽调度方法。
PCT/CN2020/074968 2019-07-11 2020-02-12 一种缓存服务器的带宽调度方法及装置 WO2021004063A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP20775568.7A EP3783869A4 (en) 2019-07-11 2020-02-12 BANDWIDTH PLANNING METHOD AND DEVICE FOR A CACHE SERVER
US17/060,032 US11102286B2 (en) 2019-07-11 2020-09-30 Bandwidth scheduling method and device for cache server

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910625761.9A CN110365765B (zh) 2019-07-11 2019-07-11 一种缓存服务器的带宽调度方法及装置
CN201910625761.9 2019-07-11

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/060,032 Continuation US11102286B2 (en) 2019-07-11 2020-09-30 Bandwidth scheduling method and device for cache server

Publications (1)

Publication Number Publication Date
WO2021004063A1 true WO2021004063A1 (zh) 2021-01-14

Family

ID=68218992

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/074968 WO2021004063A1 (zh) 2019-07-11 2020-02-12 一种缓存服务器的带宽调度方法及装置

Country Status (4)

Country Link
US (1) US11102286B2 (zh)
EP (1) EP3783869A4 (zh)
CN (1) CN110365765B (zh)
WO (1) WO2021004063A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114513464A (zh) * 2021-12-31 2022-05-17 深圳市联洲国际技术有限公司 流量负载均衡调度方法、装置、设备及存储介质
CN114531377A (zh) * 2022-01-19 2022-05-24 重庆川仪自动化股份有限公司 流程工业设备数据的流量控制方法、***、设备及介质
CN116132705A (zh) * 2023-04-17 2023-05-16 北京盘腾科技有限公司 一种直播缓存量的调节方法、装置、设备及存储介质
CN116347528A (zh) * 2023-05-29 2023-06-27 天翼云科技有限公司 一种cdn覆盖资源动态调整方法、装置及电子设备
CN116366657A (zh) * 2023-05-31 2023-06-30 天翼云科技有限公司 一种缓存服务器的数据请求调度方法及***

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110493046B (zh) * 2018-02-27 2022-08-16 贵州白山云科技股份有限公司 一种cdn网络中节点服务器带宽资源的调度方法和***
CN110365765B (zh) 2019-07-11 2022-04-22 网宿科技股份有限公司 一种缓存服务器的带宽调度方法及装置
CN112947843A (zh) * 2019-12-10 2021-06-11 北京金山云网络技术有限公司 存储***的配置、调度方法、装置及电子设备
CN111913768A (zh) * 2020-05-27 2020-11-10 广州铁路职业技术学院（广州铁路机械学校） 虚拟化云桌面及其构建方法
CN112751924B (zh) * 2020-12-29 2022-09-02 北京奇艺世纪科技有限公司 一种数据推送方法、***及装置
CN116185965B (zh) * 2023-05-04 2023-08-04 联想凌拓科技有限公司 用于服务质量控制的方法、装置、设备及介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100220622A1 (en) * 2009-02-27 2010-09-02 Yottaa Inc Adaptive network with automatic scaling
CN102301338A (zh) * 2008-11-25 2011-12-28 思杰***有限公司 用于基于健康的溢出的***和方法
CN109889569A (zh) * 2019-01-03 2019-06-14 网宿科技股份有限公司 Cdn服务调度方法及***
CN110365765A (zh) * 2019-07-11 2019-10-22 网宿科技股份有限公司 一种缓存服务器的带宽调度方法及装置

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8639813B2 (en) * 2008-11-25 2014-01-28 Citrix Systems, Inc. Systems and methods for GSLB based on SSL VPN users
CN102708011B (zh) * 2012-05-11 2014-03-26 南京邮电大学 一种云计算平台自适应任务调度方法
US9674042B2 (en) * 2013-11-25 2017-06-06 Amazon Technologies, Inc. Centralized resource usage visualization service for large-scale network topologies
CN109284871A (zh) * 2018-09-30 2019-01-29 北京金山云网络技术有限公司 资源调整方法、装置和云平台

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102301338A (zh) * 2008-11-25 2011-12-28 思杰***有限公司 用于基于健康的溢出的***和方法
US20100220622A1 (en) * 2009-02-27 2010-09-02 Yottaa Inc Adaptive network with automatic scaling
CN109889569A (zh) * 2019-01-03 2019-06-14 网宿科技股份有限公司 Cdn服务调度方法及***
CN110365765A (zh) * 2019-07-11 2019-10-22 网宿科技股份有限公司 一种缓存服务器的带宽调度方法及装置

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3783869A4 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114513464A (zh) * 2021-12-31 2022-05-17 深圳市联洲国际技术有限公司 流量负载均衡调度方法、装置、设备及存储介质
CN114513464B (zh) * 2021-12-31 2024-03-29 深圳市联洲国际技术有限公司 流量负载均衡调度方法、装置、设备及存储介质
CN114531377A (zh) * 2022-01-19 2022-05-24 重庆川仪自动化股份有限公司 流程工业设备数据的流量控制方法、***、设备及介质
CN114531377B (zh) * 2022-01-19 2024-04-02 重庆川仪自动化股份有限公司 流程工业设备数据的流量控制方法、***、设备及介质
CN116132705A (zh) * 2023-04-17 2023-05-16 北京盘腾科技有限公司 一种直播缓存量的调节方法、装置、设备及存储介质
CN116132705B (zh) * 2023-04-17 2023-06-09 北京盘腾科技有限公司 一种直播缓存量的调节方法、装置、设备及存储介质
CN116347528A (zh) * 2023-05-29 2023-06-27 天翼云科技有限公司 一种cdn覆盖资源动态调整方法、装置及电子设备
CN116347528B (zh) * 2023-05-29 2023-07-28 天翼云科技有限公司 一种cdn覆盖资源动态调整方法、装置及电子设备
CN116366657A (zh) * 2023-05-31 2023-06-30 天翼云科技有限公司 一种缓存服务器的数据请求调度方法及***

Also Published As

Publication number Publication date
EP3783869A4 (en) 2021-06-23
US11102286B2 (en) 2021-08-24
EP3783869A1 (en) 2021-02-24
CN110365765A (zh) 2019-10-22
US20210037089A1 (en) 2021-02-04
CN110365765B (zh) 2022-04-22

Similar Documents

Publication Publication Date Title
WO2021004063A1 (zh) 一种缓存服务器的带宽调度方法及装置
US8180922B2 (en) Load balancing mechanism using resource availability profiles
US8626890B2 (en) Connection pool use of runtime load balancing service performance advisories
WO2021012663A1 (zh) 一种访问日志的处理方法及装置
US7467291B1 (en) System and method for calibrating headroom margin
US20090034537A1 (en) Temporal affinity-based routing of workloads
CN110933139A (zh) 一种解决Web服务器高并发的***及方法
US20160226736A1 (en) Service call graphs for website performance
US20200012602A1 (en) Cache allocation method, and apparatus
US9501326B2 (en) Processing control system, processing control method, and processing control program
US9893973B2 (en) Real-time, low memory estimation of unique client computers communicating with a server computer
WO2016172848A1 (zh) 一种调整云计算中虚拟资源的方法及装置
WO2024094104A1 (zh) 一种动态反馈加权云存储资源调度方法、装置及设备
US7437459B2 (en) Calculation of service performance grades in a multi-node environment that hosts the services
CN109032800A (zh) 一种负载均衡调度方法、负载均衡器、服务器及***
CN117155942A (zh) 一种微服务动态自适应客户端负载均衡方法及***
US20140181174A1 (en) Distributed processing of stream data on an event protocol
US20160225043A1 (en) Determining a cost of an application
CN110569170A (zh) 服务器利用率的评价方法、装置、设备及其存储介质
CN113543160B (zh) 5g切片资源配置方法、装置、计算设备及计算机存储介质
CN115314500B (zh) 基于改进topsis模型的动态负载均衡方法
CN108471387B (zh) 一种日志流量分散控制方法及***
US20210083947A1 (en) Method and apparatus for recognizing surge in bandwidth demand
CN109639459A (zh) 一种额定带宽的调整方法、终端和可存储介质
CN106210120B (zh) 一种服务器的推荐方法及其装置

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2020775568

Country of ref document: EP

Effective date: 20200930

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20775568

Country of ref document: EP

Kind code of ref document: A1