WO2020224024A1 - 一种资源服务器的负载调节方法及系统、设备 - Google Patents

一种资源服务器的负载调节方法及系统、设备 Download PDF

Info

Publication number
WO2020224024A1
WO2020224024A1 PCT/CN2019/090320 CN2019090320W WO2020224024A1 WO 2020224024 A1 WO2020224024 A1 WO 2020224024A1 CN 2019090320 W CN2019090320 W CN 2019090320W WO 2020224024 A1 WO2020224024 A1 WO 2020224024A1
Authority
WO
WIPO (PCT)
Prior art keywords
resource server
value
operating data
current detection
load weight
Prior art date
Application number
PCT/CN2019/090320
Other languages
English (en)
French (fr)
Inventor
欧阳德志
宋本利
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 EP19874757.8A priority Critical patent/EP3761609B1/en
Priority to US16/869,141 priority patent/US11108695B2/en
Publication of WO2020224024A1 publication Critical patent/WO2020224024A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/12—Avoiding congestion; Recovering from congestion
    • H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
    • 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
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/06—Generation of reports
    • H04L43/062—Generation of reports related to network traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0805—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability
    • H04L43/0811—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability by checking connectivity
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/16—Threshold monitoring
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/29—Flow control; Congestion control using a combination of thresholds
    • 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
    • H04L47/78—Architectures of resource allocation
    • H04L47/783—Distributed allocation of resources, e.g. bandwidth brokers
    • 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
    • H04L67/1004—Server selection for load balancing
    • H04L67/1025—Dynamic adaptation of the criteria on which the server selection is based
    • 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
    • H04L67/60—Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources

Definitions

  • This application relates to the field of Internet technology, and in particular to a load adjustment method, system, and equipment for a resource server.
  • the domain name resolution server can resolve the domain name carried in the domain name resolution request into the corresponding IP address. By feeding back the resolved IP address to the user client, the user client can access the resource server at the IP address. Because the amount of resources under the same domain name may be large, multiple resource servers may exist under the same domain name to share the access traffic of many user clients.
  • the load weight value can be set for each resource server in advance, and then after receiving the access traffic of the user client, the access traffic can be distributed among the resource servers according to the load weight value. In this way, the processing performance of each resource server can be maximized.
  • the load weight value set for each resource server is fixed.
  • the actual performance of the resource server will fluctuate at any time.
  • the domain name resolution server will still assign the original access traffic to the resource server according to a fixed load weight value, which will cause the resource server to be unable to process the access assigned to itself in a timely and effective manner Traffic, even downtime due to high load. Therefore, the prior art has the problem that the overall system stability is not high enough.
  • the purpose of this application is to provide a load adjustment method, system, and device for a resource server, which can improve the overall stability of the system.
  • one aspect of the present application provides a load adjustment method for a resource server.
  • the method includes: determining that the resource server is operating in the resource server according to the historical feature value of the resource server and the operating data recorded in the current detection period. The quality floating value in the current detection period; read the load weight value currently used by the resource server, and adjust the load weight value currently used by the resource server based on the determined quality floating value, and set the adjusted The load weight value is used as the load weight value of the resource server after the current detection period ends.
  • another aspect of the present application also provides a load adjustment system for a resource server.
  • the system includes: a quality floating value determining unit for determining the historical characteristic value of the resource server and the operation recorded in the current detection period. Data to determine the quality floating value of the resource server in the current detection period; a load weight value adjustment unit for reading the load weight value currently adopted by the resource server and based on the determined quality floating value pair The load weight value currently adopted by the resource server is adjusted, and the adjusted load weight value is used as the load weight value of the resource server after the current detection period ends.
  • another aspect of the present application also provides a load regulation device for a resource server.
  • the device includes a processor and a memory.
  • the memory is used to store a computer program that is executed by the processor. , To implement the load adjustment method of the resource server described above.
  • another aspect of the present application also provides a computer storage medium for storing a computer program, which when executed by a processor, implements the foregoing resource server load adjustment method.
  • the technical solution provided by this application can perform multiple detection cycles for the recorded operating data.
  • the historical characteristic value of the resource server can be calculated, and the historical characteristic value can be used to characterize the service performance of the resource server so far.
  • the quality floating value of the resource server in the current detection period can be determined. The quality floating value can be used as a basis for adjusting the load weight value of the resource server.
  • the load weight value of the resource server after the previous detection period ends can be read, and the load weight value may be the load weight value currently adopted by the resource server.
  • the load weight value currently used by the resource server can be dynamically adjusted according to the calculated quality floating value, so that the adjusted load weight value matches the current processing performance of the resource server.
  • the load weight value currently used by the resource server can be dynamically adjusted according to the calculated quality floating value, so that the adjusted load weight value matches the current processing performance of the resource server.
  • FIG. 1 is a schematic diagram of a load adjustment method of a resource server in an embodiment of the present application
  • FIG. 2 is a flowchart of a load adjustment method of a resource server in an embodiment of the present application
  • FIG. 3 is a schematic diagram of calculating historical feature values in the current detection period in an embodiment of the present application.
  • FIG. 4 is a schematic diagram of calculation of mass floating value in an embodiment of the present application.
  • Fig. 5 is a schematic diagram of functional modules of a load adjustment system of a resource server in an embodiment of the present application.
  • This application provides a load adjustment method for a resource server.
  • the method can be applied to a GTM (Global Traffic Manager, WAN traffic management) scheduling system.
  • GTM scheduling system can perform quality monitoring for multiple resource servers under a target domain name. And log collection, and dynamically adjust the load weight of these resource servers.
  • the method may include the following steps.
  • S1 Determine the quality floating value of the resource server in the current detection period according to the historical feature value of the resource server and the operating data recorded in the current detection period.
  • the configuration information of each resource server under the target domain name can be created in advance, and the operating data of the resource server can be recorded according to the configuration information, and based on the recorded operating data, the current detection period can be calculated.
  • the historical characteristic value of the resource server can be calculated.
  • the target domain name can be any domain name that needs to dynamically adjust the load weight value.
  • the target domain name can have multiple resource servers, and these resource servers share the access traffic directed to the target domain name.
  • the configuration information of each resource server under the target domain name can be created in advance.
  • the configuration information may include multiple contents.
  • the configuration information may include scheduling reference items, the IP address of each resource server, the initial load weight value of each resource server, the monitoring and detection method of each resource server, and the information of each resource server. Monitor the detection frequency and the detection frequency for each resource server.
  • the scheduling reference item may be a characteristic value used to determine the quality of service of the resource server, and the characteristic value may be, for example, a ratio of error codes, a packet loss rate, and a time delay.
  • the initial load weight value may be generated according to the performance index of each resource server during initialization.
  • the load weight value in this application may be a numerical value or a proportional value.
  • the load weight values of these three resource servers can be 10, 20, and 20 respectively.
  • the size of the load weight value can represent the performance of the resource server.
  • the distribution ratio of the access traffic among the three resource servers can be determined.
  • the load weight ratios of the three resource servers can be calculated to be 20%, 40%, and 40%, respectively.
  • the load weight value can also be directly expressed as a proportional value. For example, the aforementioned 20%, 40%, and 40% can be directly used as the load weight value.
  • the monitoring detection mode may include a TCP detection mode, an HTTP detection mode, or a PING detection mode.
  • the monitoring detection frequency may be the frequency at which the resource server is detected using the monitoring detection method described above.
  • the detection frequency may be the result of analyzing the aforementioned monitoring detection and the frequency of dynamically adjusting the load weight value of the resource server.
  • the GTM scheduling system can start the process of dynamically adjusting the load weight value multiple times according to the detection frequency. Each time the process of adjusting the load weight value can be used as a detection cycle.
  • the detection period usually lasts for a period of time, and then the performance of the resource server during this period can be evaluated, and the load weight value can be adjusted according to the evaluation result. After completing the adjustment of the load weight value, the current detection cycle will end.
  • the operating data of the resource server can be recorded based on the configuration information.
  • the operating data may include two aspects of data.
  • the generated monitoring task may include various information such as the detection target, the detection mode, the detection port, and the source of the detection target.
  • the monitoring result may include various data such as packet loss rate, status code, number of connections, and delay.
  • the GTM scheduling system can also collect the operation log of the resource server located at the communication address according to the communication address defined by the configuration information.
  • the communication address can be the IP address of the resource server recorded in the configuration information.
  • the GTM scheduling system can collect the running log generated by the resource server during operation.
  • the running log may also include the aforementioned packet loss rate, status code, number of connections, delay and other data.
  • the recorded monitoring result and the collected running log can be used as the running data of the resource server.
  • the subsequent adjustment process of the load weight value can be performed based on these operating data.
  • a part of the recorded operating data may be invalid data or data representing that the resource server is unavailable. In practical applications, this part of the data needs to be eliminated to accurately adjust the load weight value of the resource server.
  • the data that characterizes the unavailability of the resource server can appear in many aspects such as packet loss rate, connection status, status code, and delay.
  • packet loss rate For example, for PING monitoring, data with a packet loss rate equal to 100 can be used as data that characterizes the unavailability of the resource server.
  • other status codes other than normal status codes (such as 200 to 3XX status codes) and user-specified status codes can be used as abnormal status codes, and data carrying abnormal status codes can be used as characterization resources Server unavailable data.
  • data whose delay exceeds the upper limit of the delay specified by the user can also be used as data to characterize the unavailability of the resource server.
  • the operating data that characterizes the resource unavailability of the resource server can be identified, and the operating data that characterizes the resource unavailability of the resource server can be excluded from the recorded operating data.
  • the load weight value can be adjusted based on the remaining data.
  • the GTM scheduling system can calculate the historical feature value of the resource server in the current detection cycle based on the recorded operating data.
  • the historical feature value can be regarded as before the current detection cycle (including the current cycle), the resource The average quality parameter of the server.
  • the sum of the characteristic values of each piece of operating data recorded in the current detection period can be calculated first.
  • the characteristic value may be parameters such as packet loss rate, time delay, and error code ratio. Taking time delay as an example, the sum of the characteristic values may represent the sum of time delays characterized by various pieces of operating data in the current detection period.
  • the characteristic value of each piece of operating data in the current detection period may be accumulated, and the accumulated result may be used as the sum of the characteristic value of each piece of operating data in the current detection period.
  • the characteristic value of each piece of operating data in the current detection cycle can be limited. Specifically, if the characteristic value of the current operation data is greater than the preset threshold value, the characteristic value of the current operation data may be modified to the preset threshold value.
  • the preset threshold is a time delay of 20ms. If the time delay represented by a certain operating data in the current detection period is 30ms, the time delay represented by the operating data can be corrected to 20ms. The latter result is used to calculate the sum of the total delay in the current detection period.
  • the historical characteristic value of the resource server after the end of the last detection period can be read, and the historical characteristic value can be used as the average quality parameter of the resource server characterized by other running data in addition to the running data in the current detection period.
  • the sum of the characteristic value of the historical operation data can be calculated. It should be noted that the data volume of historical operation data before the current detection period does not include the data volume of the operation data in the current detection period. In this way, when calculating the sum of the characteristic values of the historical operation data, the product of the read historical characteristic values and the data amount of the historical operation data may be used as the sum of the characteristic values of the historical operation data.
  • the sum of the characteristic values of the historical operating data can represent the sum of the time delays represented by various pieces of historical operating data.
  • the historical characteristic value of the resource server in the current detection period may be calculated according to the sum of the characteristic values calculated in the current detection period and the sum of the characteristic values of the historical operation data. Specifically, the total data volume of the operating data recorded in the current inspection period and the historical operating data can be calculated first, and then the sum of the characteristic values calculated in the current inspection period and the historical operating data can be accumulated. The sum of the characteristic values of, and the ratio of the cumulative result to the total data amount is used as the historical characteristic value of the resource server in the current detection period.
  • the historical feature value of the resource server in the current detection period can be expressed by the following formula:
  • h t represents the historical feature value in the current detection cycle
  • h t-1 represents the historical feature value after the end of the previous detection cycle
  • hc can represent all historical operating data including the operating data of the current detection cycle
  • Cc represents the data volume of the operating data in the current detection cycle.
  • hc-cc can represent the data volume of the historical operating data before the current detection cycle (not including the data volume of the current detection cycle)
  • c k represents the characteristic value of the k-th operating data in the current detection cycle
  • ⁇ c k can represent the sum of the characteristic values of the operating data in the current detection cycle.
  • the average quality parameter of the resource server before the current detection period can be obtained.
  • the average quality parameter can then be the aforementioned historical feature value in the current detection period.
  • the historical feature value obtained by the above calculation can be used as an average quality parameter of all historical operating data including the current detection period. Therefore, the actual characteristic value represented by each piece of operating data in the current detection period can be compared with the historical characteristic value, so as to determine whether the performance of the resource server exhibits large fluctuations in the current detection period.
  • the first operating data whose characteristic value is greater than the historical characteristic value can be identified, and the characteristic value can be identified as being less than or equal to the historical characteristic. Value of the second running data.
  • the first operating data and the second operating data can respectively represent different types of data.
  • the first operating data and the second operating data may respectively represent different service qualities of the resource server.
  • the historical feature value can be a historical average value of time delay. If the time delay is greater than the historical average value of time delay in the running data recorded in the current detection period, it indicates that the resource server has poor service quality .
  • the quality floating value can be calculated separately for the first operating data and the second operating data. Since the first operating data and the second operating data represent the quality of service respectively, the two calculated Each quality floating value is usually one positive and one negative. In this case, the sum of the calculated quality floating values can be used as the quality floating value of the resource server in the current detection period. In other words, the operating data that characterizes the poor service quality of the resource server will cause the quality floating value to tend to a negative value, while the operating data that characterizes the resource server's good quality of service will cause the quality floating value to tend to a positive value.
  • the quality fluctuation value in the current detection period can be expressed by the following formula:
  • S represents the quality floating value in the current detection period
  • y i represents the characteristic value represented by the i-th data in the first operating data
  • x j represents the characteristic value represented by the j-th data in the second operating data.
  • U represents the preset threshold used to modify the characteristic value of the current operating data
  • datacount represents the total amount of data in the current detection cycle
  • m represents the total amount of data in the first operating data
  • n represents the total amount of data in the second operating data the amount.
  • It can represent the quality floating value of the first running data, It can represent the quality floating value of the second running data.
  • the quality floating value in the current detection period can be compared with the designated floating interval, so as to adjust the load weight value currently used by the resource server accordingly.
  • the load weight value currently used by the resource server may be the load weight value adjusted after the last detection period ends. If there is no previous detection period, then the load weight value currently used by the resource server can be Is the initial load weight value in the configuration information.
  • the determined quality floating value is within the designated floating interval, it means that the quality fluctuation in the current detection period is allowed. At this time, the load weight value currently used by the resource server can be kept unchanged. change. If the determined quality floating value is greater than the upper limit of the designated floating interval, it means that the service quality of the resource server has improved greatly in the current detection period. At this time, the resource server has higher processing performance. Therefore, you can follow
  • the preset adjustment amplitude increases the load weight value currently adopted by the resource server.
  • the preset adjustment amplitude may be a preset fixed value, and each time the load weight value needs to be adjusted, the preset adjustment amplitude is used for adjustment by one adjustment unit.
  • the preset adjustment amplitude reduces the load weight value currently adopted by the resource server.
  • the adjusted load weight value may be used as the load weight value of the resource server after the current detection period ends. Specifically, if the adjusted load weight value is within the specified load interval, it means that the adjusted load weight value is reasonable. Therefore, the adjusted load weight value can be used as the resource server's value after the current detection period ends. Load weight value. However, if the adjusted load weight value is greater than the upper limit of the specified load interval, it means that the load weight value adjusted according to the above preset adjustment amplitude is too large. In this case, the specified load interval can be The upper limit value is used as the load weight value of the resource server after the current detection period ends.
  • the adjusted load weight value is less than the lower limit value of the specified load interval, it means that the load weight value adjusted according to the above preset adjustment amplitude is too small. At this time, the lower limit of the specified load interval can be adjusted.
  • the limit value is used as the load weight value of the resource server after the current detection period ends.
  • the load weight value of the current resource server can be kept unchanged, and the load weight value of other resource servers except the current resource server under the target domain name can be increased. In this way, the load weight value of the current resource server is reduced in disguise, so that more access traffic can be distributed to other resource servers.
  • the GTM scheduling system may send the scheduling policy characterizing the adjusted load weight value to the DNS authoritative server to This allows the DNS authoritative server to distribute the access request directed to the target domain name among the resource servers according to the respective load weight values characterized by the scheduling policy.
  • the DNS authoritative server can dynamically change the allocation policy of access requests after each detection cycle, so that the number of access requests processed by the resource server matches its processing performance.
  • this application also provides a load adjustment system for a resource server, the system including:
  • the quality floating value determining unit is configured to determine the quality floating value of the resource server in the current detection period according to the historical characteristic value of the resource server and the operating data recorded in the current detection period;
  • the load weight value adjustment unit is configured to read the load weight value currently used by the resource server, and adjust the load weight value currently used by the resource server based on the determined quality floating value, and adjust the adjusted load The weight value is used as the load weight value of the resource server after the current detection period ends.
  • system further includes:
  • the monitoring result recording unit is configured to generate a monitoring task for the resource server according to the monitoring detection mode and monitoring detection frequency, and record the monitoring result corresponding to the resource server when the monitoring task is executed;
  • the operation log collection unit is configured to collect the operation log of the resource server located at the communication address according to the communication address;
  • the operating data determining unit is configured to use the recorded monitoring results and collected operating logs as operating data of the resource server.
  • system further includes:
  • the first elimination unit is configured to identify operating data carrying invalid status codes from the operating data, and remove the operating data carrying invalid status codes from the recorded operating data;
  • the second elimination unit is used to identify the operating data that characterizes the resource unavailability of the resource server from the operating data, and remove the operating data that characterizes the resource unavailability of the resource server from the recorded operating data .
  • system further includes a historical feature value calculation unit, and the historical feature value calculation unit includes:
  • the current detection period calculation module is used to calculate the sum of the characteristic values of each piece of operating data recorded in the current detection period
  • the historical operating data calculation module is used to read the historical feature value of the resource server after the last detection period ends, and according to the read historical feature value and the data of the historical operating data before the current detection period Calculate the sum of the characteristic values of the historical operating data;
  • the historical characteristic value calculation module is configured to calculate the historical characteristic value of the resource server in the current detection period according to the sum of the characteristic values calculated in the current detection period and the sum of the characteristic values of the historical operation data.
  • the current detection period calculation module includes:
  • the accumulation module is used to accumulate the characteristic values of each piece of operating data in the current inspection period, and use the accumulated result as the sum of the characteristic values of each piece of operating data in the current inspection period;
  • the characteristic value is greater than the preset threshold, and the characteristic value of the current operating data is modified to the preset threshold.
  • the historical operating data calculation module includes:
  • the product module is used to take the product of the read historical characteristic value and the data volume of the historical operating data as the sum of the characteristic values of the historical operating data.
  • the historical feature value calculation module includes:
  • the total data volume calculation module is used to calculate the total data volume of the operating data recorded in the current detection period and the historical operating data;
  • the ratio calculation module is used to accumulate the sum of the characteristic values calculated in the current detection period and the sum of the characteristic values of the historical operating data, and use the ratio of the accumulated result to the total data amount as the current The historical characteristic value of the resource server in the detection period.
  • the quality floating value determination unit includes:
  • the operating data identification module is used for identifying the first operating data whose characteristic value is greater than the historical characteristic value among the operating data recorded in the current detection period, and identifying the characteristic value less than or equal to the historical characteristic value Second operating data;
  • the quality floating value calculation module is configured to calculate the quality floating values for the first operating data and the second operating data respectively, and use the sum of the calculated quality floating values as the current detection by the resource server The floating value of the mass during the period.
  • the load weight value adjustment unit includes:
  • a maintaining module configured to keep the load weight value currently used by the resource server unchanged if the determined quality floating value is within a designated floating interval
  • An increase module configured to increase the load weight value currently adopted by the resource server according to a preset adjustment amplitude if the determined quality floating value is greater than the upper limit of the designated floating interval;
  • the reduction module is configured to reduce the load weight value currently adopted by the resource server according to the preset adjustment amplitude if the determined quality floating value is less than the lower limit of the designated floating interval.
  • the load weight value adjustment unit includes:
  • a first determining module configured to, if the adjusted load weight value is within a specified load interval, use the adjusted load weight value as the load weight value of the resource server after the current detection period ends;
  • the second determining module is configured to, if the adjusted load weight value is greater than the upper limit of the specified load interval, use the upper limit of the specified load interval as the load of the resource server after the current detection period ends.
  • the third determining module is configured to, if the adjusted load weight value is less than the lower limit of the specified load interval, use the lower limit of the specified load interval as the load of the resource server after the current detection period ends. Weights.
  • the load weight value adjustment unit includes:
  • a drop control module configured to keep the load weight value of the current resource server unchanged if the drop represented by the quality floating value of the current resource server reaches the preset drop threshold, and increase the target domain name to remove the current resource server Load weight value of other resource servers.
  • system further includes:
  • the scheduling policy issuing unit is configured to, after the load weight value adjustment is completed for each resource server under the target domain name, send the scheduling policy representing the adjusted load weight value to the DNS authoritative server, so that the DNS is authoritative
  • the server distributes the access request directed to the target domain name among the resource servers according to each load weight value represented by the scheduling policy.
  • the present application also provides a load adjustment device for a resource server.
  • the device includes a processor and a memory.
  • the memory is used to store a computer program.
  • the resource server can be Load adjustment method.
  • the present application also provides a computer storage medium, which is used to store a computer program, and when the computer program is executed by a processor, the aforementioned resource server load adjustment method can be implemented.
  • the memory and computer storage medium may include a physical device for storing information, and the information is usually digitized and then stored in a medium using electrical, magnetic, or optical methods.
  • the memory and computer storage medium described in this embodiment may also include: devices that use electrical energy to store information, such as RAM, ROM, etc.; devices that use magnetic energy to store information, such as hard disks, floppy disks, magnetic tapes, magnetic core memories, and magnetic bubbles Memory, U disk; a device that uses optical methods to store information, such as CD or DVD.
  • devices that use electrical energy to store information such as RAM, ROM, etc.
  • devices that use magnetic energy to store information such as hard disks, floppy disks, magnetic tapes, magnetic core memories, and magnetic bubbles Memory, U disk
  • a device that uses optical methods to store information such as CD or DVD.
  • quantum memory graphene memory and so on.
  • the processor may be implemented in any suitable manner.
  • the processor may take the form of, for example, a microprocessor or a processor and a computer-readable medium storing computer-readable program codes (for example, software or firmware) executable by the (micro)processor, logic gates, switches, special-purpose integrated Circuit (Application Specific Integrated Circuit, ASIC), programmable logic controller and embedded microcontroller form, etc.
  • program codes for example, software or firmware
  • the technical solution provided by this application can perform multiple detection cycles for the recorded operating data.
  • the historical characteristic value of the resource server can be calculated, and the historical characteristic value can be used to characterize the service performance of the resource server so far.
  • the quality floating value of the resource server in the current detection period can be determined. The quality floating value can be used as a basis for adjusting the load weight value of the resource server.
  • the load weight value of the resource server after the previous detection period ends can be read, and the load weight value may be the load weight value currently adopted by the resource server.
  • the load weight value currently used by the resource server can be dynamically adjusted according to the calculated quality floating value, so that the adjusted load weight value matches the current processing performance of the resource server.
  • the load weight value currently used by the resource server can be dynamically adjusted according to the calculated quality floating value, so that the adjusted load weight value matches the current processing performance of the resource server.
  • each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware.
  • the above technical solutions can be embodied in the form of software products, which can be stored in computer-readable storage media, such as ROM/RAM, magnetic A disc, an optical disc, etc., include a number of instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiment.

Abstract

本申请公开了一种资源服务器的负载调节方法及系统、设备，其中，所述方法应用于调度系统中，所述方法包括：根据资源服务器的历史特征值，以及当前检测周期内记录的运行数据，确定所述资源服务器在所述当前检测周期内的质量浮动值；读取所述资源服务器当前采用的负载权重值，并基于确定的所述质量浮动值对所述资源服务器当前采用的负载权重值进行调整，并将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值。本申请提供的技术方案，能够提高系统的整体稳定度。

Description

一种资源服务器的负载调节方法及系统、设备
交叉引用
本申请引用于2019年05月09日递交的名称为“一种资源服务器的负载调节方法及系统、设备”的第201910383394.6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，特别涉及一种资源服务器的负载调节方法及系统、设备。
背景技术
在域名解析系统中，域名解析服务器在接收到用户客户端发来的域名解析请求后，可以将该域名解析请求中携带的域名，解析为对应的IP地址。通过将解析得到的IP地址反馈给用户客户端，便可以使得用户客户端访问IP地址处的资源服务器。由于同一个域名下的资源量可能较多，因此同一个域名下可以存在多个资源服务器，以分担众多用户客户端的访问流量。
目前，可以预先给各个资源服务器设置负载权重值，然后在接收到用户客户端的访问流量后，可以将访问流量按照负载权重值在各个资源服务器之间分配。这样，可以最大化地利用各个资源服务器的处理性能。
现有的这种方式，给每个资源服务器设置的负载权重值都是固定的。在实际应用中，资源服务器的实际性能会随时发生波动。然而，即使某个资源服务器的处理性能下降，域名解析服务器还是会按照固定的负载权重值给该资源服务器分配原有的访问流量，这会导致该资源服务器无法及时有效地处理分配至自身的访问流量，甚至会由于负载过高而宕机。因此，现有技术会存在系统整体稳定度不够高的问题。
发明内容
本申请的目的在于提供一种资源服务器的负载调节方法及系统、设备，能够提高系统的整体稳定度。
为实现上述目的，本申请一方面提供一种资源服务器的负载调节方法，所述方法包括：根据资源服务器的历史特征值，以及当前检测周期内记录的运行数据，确定所述资源服务器在所述当前检测周期内的质量浮动值；读取所述资源服务器当前采用的负载权重值，并基于确定的所述质量浮动值对所述资源服务器当前采用的负载权重值进行调整，并将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值。
为实现上述目的，本申请另一方面还提供一种资源服务器的负载调节系统，所述系统包括：质量浮动值确定单元，用于根据资源服务器的历史特征值，以及当前检测周期内记录的运行数据，确定所述资源服务器在所述当前检测周期内的质量浮动值；负载权重值调整单元，用于读取所述资源服务器当前采用的负载权重值，并基于确定的所述质量浮动值对所述资源服务器当前采用的负载权重值进行调整，并将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值。
为实现上述目的，本申请另一方面还提供一种资源服务器的负载调节设备，所述设备包括处理器和存储器，所述存储器用于存储计算机程序，所述计算机程序被所述处理器执行时，实现上述的资源服务器的负载调节方法。
为实现上述目的，本申请另一方面还提供一种计算机存储介质，所述计算机存储介质用于存储计算机程序，所述计算机程序被处理器执行时，实现上述的资源服务器的负载调节方法。
由上可见，本申请提供的技术方案，针对记录的运行数据，可以进行多个检测周期的检测。在当前检测周期内，可以计算出资源服务器的历史特征值，该历史特征值可以用于表征资源服务器截至目前为止的服务性能。基于该历史特征值和当前检测周期内记录的运行数据，可以确定该资源服务器在当前检测周期内的质量浮动值。该质量浮动值可以作为该资源服务器的负载权重值的调整依据。后续，可以读取该资源服务器在上一个检测周期结束后的负载权重值，该负载权重值便可以是资源服务器当前采用的负载权重值。然后，可以根据上述计算出的质量浮动值，对该资源服务器当前采用的负载权重值进行动态调整，以使得调整后的负载权重值，与该资源服务器当前的处理性能相匹配。这样， 当资源服务器当前的处理性能下降时，分配至该资源服务器的访问流量也会适当减少；而当资源服务器当前的处理性能提升时，分配至该资源服务器的访问流量也会适当增多。因此，通过动态调节资源服务器的负载权重值，能够提升系统的整体稳定性。
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他的附图。
图1是本申请实施例中资源服务器的负载调节方法示意图；
图2是本申请实施例中资源服务器的负载调节方法流程图；
图3是本申请实施例中当前检测周期内的历史特征值的计算示意图；
图4是本申请实施例中质量浮动值的计算示意图；
图5是本申请实施例中资源服务器的负载调节系统的功能模块示意图。
具体实施例
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例作进一步地详细描述。
本申请提供一种资源服务器的负载调节方法，所述方法可以应用于GTM(Global Traffic Manager，广域网流量管理)调度系统中，所述GTM调度系统可以针对目标域名下的多个资源服务器进行质量监控和日志采集，并动态地对这些资源服务器的负载权重值进行调整。
具体地，请参阅图1和图2，所述方法可以包括以下步骤。
S1：根据资源服务器的历史特征值，以及当前检测周期内记录的运行数据，确定所述资源服务器在所述当前检测周期内的质量浮动值。
在一个实施例中，可以预先创建目标域名下各个资源服务器的配置信息，并根据所述配置信息，记录所述资源服务器的运行数据，并基于记录的所述运行数据，计算当前检测周期内所述资源服务器的历史特征值。
在本实施例中，目标域名可以是任一需要进行负载权重值动态调节的域 名，该目标域名下可以具备多个资源服务器，这些资源服务器共同分担指向该目标域名的访问流量。
在所述GTM调度系统中，可以预先创建该目标域名下的各个资源服务器的配置信息。所述配置信息可以包含多项内容，例如，所述配置信息可以包含调度参考项、各个资源服务器的IP地址、各个资源服务器的初始负载权重值、各个资源服务器的监控探测方式、各个资源服务器的监控探测频率以及针对各个资源服务器的检测频率等。其中，所述调度参考项可以是用于判断资源服务器的服务质量的特征值，该特征值例如可以是错误码的比例、丢包率、时延等。所述初始负载权重值可以根据各个资源服务器初始化时的性能指标生成。
需要说明的是，在本申请中的负载权重值，可以是一个数值，也可以是一个比例值。例如，当前共有3个资源服务器，这3个资源服务器的负载权重值可以分别为10、20、20，负载权重值的大小，可以表征资源服务器的性能高低。后续，通过分别计算各个资源服务器的负载权重值在负载权重总值中的比例，从而可以决定访问流量在这3个资源服务器之间的分配比例。例如，按照上述的负载权重值，可以计算出这3个资源服务器的负载权重比例分别为20％、40％以及40％。此外，所述负载权重值还可以直接表示为比例值。例如，上述的20％、40％和40％便可以直接作为负载权重值。
上述的监控探测方式，可以根据实际情况灵活调整。例如，所述监控探测方式可以包括TCP探测方式、HTTP探测方式或者PING探测方式等。所述监控探测频率，可以是采用上述监控探测方式对资源服务器进行探测的频率。所述检测频率，可以是分析上述的监控探测的结果，以及对资源服务器的负载权重值进行动态调整的频率。在本实施例中，GTM调度系统可以按照所述检测频率，开启多次动态调整负载权重值的过程。每次调整负载权重值的过程，都可以作为一个检测周期。该检测周期通常会持续一段时间，然后可以对这段时间内资源服务器的性能进行评估，并根据评估结果进行负载权重值的调整。在完成负载权重值的调整之后，便会结束当前的检测周期。
在本实施例中，对目标域名下的各个资源服务器创建了配置信息后，便可以基于该配置信息，记录所述资源服务器的运行数据。在实际应用中，所述运行数据可以包括两方面的数据。一方面，可以根据所述配置信息限定的监控探测方式和监控探测频率，为所述资源服务器生成监控任务，并在所述监控任 务执行时，记录所述资源服务器对应的监控结果。具体地，生成的所述监控任务中，可以包含探测目标、探测方式、探测的端口以及探测目标的来源等各项信息。其中，该监控结果例如可以包括丢包率、状态码、连接数、时延等各项数据。另一方面，GTM调度系统还可以根据所述配置信息限定的通信地址，采集位于所述通信地址处的资源服务器的运行日志。该通信地址便可以是配置信息中记录的资源服务器的IP地址，通过位于该IP地址的资源服务器提供的访问接口，GTM调度系统可以采集到资源服务器在运行时产生的运行日志。该运行日志中也可以包含上述的丢包率、状态码、连接数、时延等各项数据。
这样，通过监控和日志采集，可以将记录的所述监控结果和采集的所述运行日志作为所述资源服务器的运行数据。后续可以基于这些运行数据进行负载权重值的调整过程。
在一个实施例中，记录的所述运行数据中，可能有一部分是无效数据或者表征资源服务器不可用的数据。在实际应用中，需要将这部分数据剔除，才能准确地对资源服务器的负载权重值进行调整。具体地，上述需要剔除的数据通常具备一定的特征。例如，对于PING监控、TCP监控以及HTTP监控而言，无效数据中通常会携带无效状态码，该无效状态码在实际应用中例如可以是0,2,3,5,7,9，这些无效状态码可以通过CODE！＝(0,2,3,5,7,9)这样的形式来表示。因此，在记录了所述运行数据后，可以从所述运行数据中，识别出携带无效状态码的运行数据，并将所述携带无效状态码的运行数据从记录的运行数据中剔除。此外，表征资源服务器不可用的数据，可以出现于丢包率、连接状态、状态码、时延等多个方面。例如，对于PING监控而言，可以将丢包率等于100的数据作为表征资源服务器不可用的数据。对于TCP监控而言，可以将TCP连接状态为1(state＝1)的数据作为表征资源服务器不可用的数据。对于HTTP监控而言，可以将除正常状态码(例如200至3XX的状态码)和用户指定的状态码之外的其它状态码作为异常状态码，并可以将携带异常状态码的数据作为表征资源服务器不可用的数据。另外，对于HTTP监控而言，还可以将时延超出用户指定的时延上限的数据，作为表征资源服务器不可用的数据。这样，可以从所述运行数据中，识别出表征所述资源服务器的资源不可用的运行数据，并将表征所述资源服务器的资源不可用的运行数据从记录的运行数据中剔除。
当然，上述例举的各种情况只是为了更好地理解本申请的方案，并不代 表本申请的方案仅适应于上述例举的情况。
在本实施例中，将无效数据和表征资源服务器不可用的数据剔除后，可以基于剩余的数据进行负载权重值的调整。
具体地，GTM调度系统可以基于记录的所述运行数据，计算当前检测周期内所述资源服务器的历史特征值，所述历史特征值，可以作为在当前检测周期以前(包含当前周期)，该资源服务器的平均质量参数。请参阅图3，在实际应用中，首先可以计算所述当前检测周期内记录的各条运行数据的特征值之和。所述特征值可以是丢包率、时延、错误码的比例等参数。以时延为例，所述特征值之和，可以代表当前检测周期内各条运行数据所表征的时延之和。具体地，可以累计所述当前检测周期内，各条运行数据的特征值，并将累计的结果作为所述当前检测周期内各条运行数据的特征值之和。
需要说明的是，在实际应用中，可以对当前检测周期内各条运行数据的特征值进行限定。具体地，若当前运行数据的特征值大于预设阈值，则可以将所述当前运行数据的特征值修改为所述预设阈值。举例来说，所述预设阈值为20ms的时延，那么如果当前检测周期内的某个运行数据表征的时延为30ms，那么可以将该运行数据表征的时延修正为20ms，并根据修正后的结果来计算当前检测周期内总的时延之和。
然后，可以读取上一个检测周期结束后所述资源服务器的历史特征值，该历史特征值可以作为除当前检测周期内的运行数据之外的，其它运行数据所表征的资源服务器的平均质量参数。根据读取的所述历史特征值，和位于所述当前检测周期之前的历史运行数据的数据量，可以计算所述历史运行数据的特征值之和。需要说明的是，这里位于所述当前检测周期之前的历史运行数据的数据量，不包含当前检测周期内的运行数据的数据量。这样，在计算所述历史运行数据的特征值之和时，可以将读取的所述历史特征值和所述历史运行数据的数据量的乘积，作为所述历史运行数据的特征值之和。同样以时延为例，该历史运行数据的特征值之和，便可以代表各条历史运行数据所表征的时延之和。最终，可以根据所述当前检测周期内计算得到的特征值之和以及所述历史运行数据的特征值之和，计算所述当前检测周期内所述资源服务器的历史特征值。具体地，首先可以计算所述当前检测周期内记录的运行数据和所述历史运行数据的总数据量，然后，可以累计所述当前检测周期内计算得到的特征值之和与 所述历史运行数据的特征值之和，并将累计的结果与所述总数据量的比值，作为所述当前检测周期内所述资源服务器的历史特征值。
在一个应用示例中，所述资源服务器在当前检测周期内的历史特征值可以通过以下公式表示：
Figure PCTCN2019090320-appb-000001
其中，h t表示所述当前检测周期内的历史特征值，h t-1则表示上一个检测周期结束后的历史特征值，hc可以表示包含当前检测周期的运行数据在内的全部历史运行数据的数据量，cc则表示当前检测周期内的运行数据的数据量，这样，hc-cc可以表示位于所述当前检测周期之前的历史运行数据的数据量(不包含当前检测周期的数据量)，c k表示当前检测周期内第k个运行数据的特征值，∑c k则可以表示当前检测周期内运行数据的特征值之和。
这样，通过上述的计算过程，可以得到在当前检测周期以前(包含当前检测周期在内)，资源服务器的平均质量参数。该平均质量参数便可以是上述的当前检测周期内的历史特征值。
由上可见，在计算当前检测周期内的历史特征值时，也需要用到当前检测周期之前的其它历史运行数据，这样处理的目的在于，可以使得历史特征值的变化比较平滑，避免由于局部数据的特殊性，导致历史特征值的起伏较大，从而可以保证数据的准确程度。
在本实施例中，上述计算得到的历史特征值，可以作为包含当前检测周期在内的所有历史运行数据的平均质量参数。因此，可以将当前检测周期内各条运行数据表征的实际特征值，与该历史特征值进行比较，从而判断当前检测周期内，资源服务器的性能是否出现较大的起伏。
具体地，请参阅图4，可以在所述当前检测周期内记录的运行数据中，识别出特征值大于所述历史特征值的第一运行数据，以及识别出特征值小于或者等于所述历史特征值的第二运行数据。这样，第一运行数据和第二运行数据可以分别表示不同类别的数据。在实际应用中，第一运行数据和第二运行数据可以分别表示资源服务器的不同服务质量。以时延为例，所述历史特征值可以是历史的时延均值，如果当前检测周期内记录的运行数据中，时延大于该历史的时延均值，则表示资源服务器存在不好的服务质量。而如果当前检测周期内记录的运行数据中，时延小于或者等于该历史的时延均值，则表示资源服务器的 服务质量较好。此时，可以针对所述第一运行数据和所述第二运行数据，分别计算质量浮动值，由于第一运行数据和第二运行数据分别代表了服务质量的好和坏，那么计算得到的两个质量浮动值通常是一正一负，此时，可以将计算得到的质量浮动值之和，作为所述资源服务器在所述当前检测周期内的质量浮动值。也就是说，表征资源服务器的服务质量不佳的运行数据，会导致质量浮动值趋向负值，而表征资源服务器的服务质量较好的运行数据，会导致质量浮动值趋向正值。
具体地，在一个应用示例中，所述当前检测周期内的质量浮动值可以通过以下公式表示：
Figure PCTCN2019090320-appb-000002
其中，S表示所述当前检测周期内的质量浮动值，y i表示第一运行数据中的第i个数据表征的特征值，x j表示第二运行数据中的第j个数据表征的特征值，u表示用于修正当前运行数据的特征值的所述预设阈值，datacount表示当前检测周期内的数据总量，m表示第一运行数据的数据总量，n表示第二运行数据的数据总量。
上述公式中，
Figure PCTCN2019090320-appb-000003
可以表示第一运行数据的质量浮动值，
Figure PCTCN2019090320-appb-000004
可以表示第二运行数据的质量浮动值。
S3：读取所述资源服务器当前采用的负载权重值，并基于确定的所述质量浮动值对所述资源服务器当前采用的负载权重值进行调整，并将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值。
在本实施例中，在计算出当前检测周期内的质量浮动值后，可以将该质量浮动值与指定浮动区间进行对比，从而对资源服务器当前所采用的负载权重值进行相应的调整。
具体地，所述资源服务器当前所采用的负载权重值，可以是上一个检测周期结束后调整得到的负载权重值，如果不存在上一个检测周期，那么资源服务器当前所采用的负载权重值便可以是配置信息中的初始负载权重值。
在本实施例中，若确定的所述质量浮动值位于所述指定浮动区间内，则表示当前检测周期内的质量波动是允许的，此时可以保持所述资源服务器当前 采用的负载权重值不变。若确定的所述质量浮动值大于所述指定浮动区间的上限值，则表示当前检测周期内，资源服务器的服务质量提升较大，此时资源服务器具备较高的处理性能，因此，可以按照预设调整幅值提高所述资源服务器当前采用的负载权重值。所述预设调整幅值可以是预先设置的固定值，每次需要调整负载权重值时，便以该预设调整幅值为一个调整单元进行调整。若确定的所述质量浮动值小于所述指定浮动区间的下限值，则表示当前检测周期内，资源服务器的服务质量下降较大，此时资源服务器的处理性能较差，因此可以按照所述预设调整幅值减小所述资源服务器当前采用的负载权重值。
在本实施例中，在完成当前检测周期的负载权重值的调整过程后，可以将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值。具体地，若调整后的负载权重值位于指定负载区间内，则表示调整后的负载权重值比较合理，因此可以将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值。然而，若调整后的负载权重值大于所述指定负载区间的上限值，则表示按照上述的预设调整幅值调整后的负载权重值过大，此时，可以将所述指定负载区间的上限值作为所述资源服务器在所述当前检测周期结束后的负载权重值。相反，若调整后的负载权重值小于所述指定负载区间的下限值，则表示按照上述的预设调整幅值调整后的负载权重值过小，此时可以将所述指定负载区间的下限值作为所述资源服务器在所述当前检测周期结束后的负载权重值。
在一个实施例中，针对质量浮动值表征资源服务器的处理性能下降的情况，若当前的资源服务器的质量浮动值表征的跌幅达到预设跌幅阈值，在表示目标域名下的当前的资源服务器具备较差的处理性能，此时，可以保持所述当前的资源服务器的负载权重值不变，并提高目标域名下除所述当前的资源服务器之外的其它资源服务器的负载权重值。这样，变相降低了当前的资源服务器的负载权重值，从而可以将更多的访问流量分配至其它的资源服务器处。
在本实施例中，在针对所述目标域名下的各个所述资源服务器均完成负载权重值调整后，GTM调度系统可以将表征调整后的各个负载权重值的调度策略发送至DNS权威服务器，以使得所述DNS权威服务器可以按照所述调度策略表征的各个负载权重值，将指向所述目标域名的访问请求在各个所述资源服务器之间进行分配。这样，DNS权威服务器可以在每个检测周期后都动态更改访问请 求的分配策略，使得资源服务器处理的访问请求的数量与自身的处理性能相匹配。
请参阅图5，本申请还提供一种资源服务器的负载调节系统，所述系统包括：
质量浮动值确定单元，用于根据资源服务器的历史特征值，以及当前检测周期内记录的运行数据，确定所述资源服务器在所述当前检测周期内的质量浮动值；
负载权重值调整单元，用于读取所述资源服务器当前采用的负载权重值，并基于确定的所述质量浮动值对所述资源服务器当前采用的负载权重值进行调整，并将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值。
在一个实施例中，所述系统还包括：
监控结果记录单元，用于根据监控探测方式和监控探测频率，为所述资源服务器生成监控任务，并在所述监控任务执行时，记录所述资源服务器对应的监控结果；
运行日志采集单元，用于根据通信地址，采集位于所述通信地址处的资源服务器的运行日志；
运行数据确定单元，用于将记录的所述监控结果和采集的运行日志作为所述资源服务器的运行数据。
在一个实施例中，所述系统还包括：
第一剔除单元，用于从所述运行数据中，识别出携带无效状态码的运行数据，并将所述携带无效状态码的运行数据从记录的运行数据中剔除；
第二剔除单元，用于从所述运行数据中，识别出表征所述资源服务器的资源不可用的运行数据，并将表征所述资源服务器的资源不可用的运行数据从记录的运行数据中剔除。
在一个实施例中，所述系统还包括历史特征值计算单元，所述历史特征值计算单元包括：
当前检测周期计算模块，用于计算所述当前检测周期内记录的各条运行数据的特征值之和；
历史运行数据计算模块，用于读取上一个检测周期结束后所述资源服务器 的历史特征值，并根据读取的所述历史特征值，和位于所述当前检测周期之前的历史运行数据的数据量，计算所述历史运行数据的特征值之和；
历史特征值计算模块，用于根据所述当前检测周期内计算得到的特征值之和以及所述历史运行数据的特征值之和，计算所述当前检测周期内所述资源服务器的历史特征值。
在一个实施例中，所述当前检测周期计算模块包括：
累计模块，用于累计所述当前检测周期内，各条运行数据的特征值，并将累计的结果作为所述当前检测周期内各条运行数据的特征值之和；其中，若当前运行数据的特征值大于预设阈值，将所述当前运行数据的特征值修改为所述预设阈值。
在一个实施例中，所述历史运行数据计算模块包括：
乘积模块，用于将读取的所述历史特征值和所述历史运行数据的数据量的乘积，作为所述历史运行数据的特征值之和。
在一个实施例中，所述历史特征值计算模块包括：
总数据量计算模块，用于计算所述当前检测周期内记录的运行数据和所述历史运行数据的总数据量；
比值计算模块，用于累计所述当前检测周期内计算得到的特征值之和与所述历史运行数据的特征值之和，并将累计的结果与所述总数据量的比值，作为所述当前检测周期内所述资源服务器的历史特征值。
在一个实施例中，所述质量浮动值确定单元包括：
运行数据识别模块，用于在所述当前检测周期内记录的运行数据中，识别出特征值大于所述历史特征值的第一运行数据，以及识别出特征值小于或者等于所述历史特征值的第二运行数据；
质量浮动值计算模块，用于针对所述第一运行数据和所述第二运行数据，分别计算质量浮动值，并将计算得到的质量浮动值之和，作为所述资源服务器在所述当前检测周期内的质量浮动值。
在一个实施例中，所述负载权重值调整单元包括：
保持模块，用于若确定的所述质量浮动值位于指定浮动区间内，保持所述资源服务器当前采用的负载权重值不变；
提高模块，用于若确定的所述质量浮动值大于所述指定浮动区间的上限值， 按照预设调整幅值提高所述资源服务器当前采用的负载权重值；
减小模块，用于若确定的所述质量浮动值小于所述指定浮动区间的下限值，按照所述预设调整幅值减小所述资源服务器当前采用的负载权重值。
在一个实施例中，所述负载权重值调整单元包括：
第一确定模块，用于若调整后的负载权重值位于指定负载区间内，将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值；
第二确定模块，用于若调整后的负载权重值大于所述指定负载区间的上限值，将所述指定负载区间的上限值作为所述资源服务器在所述当前检测周期结束后的负载权重值；
第三确定模块，用于若调整后的负载权重值小于所述指定负载区间的下限值，将所述指定负载区间的下限值作为所述资源服务器在所述当前检测周期结束后的负载权重值。
在一个实施例中，所述负载权重值调整单元包括：
跌幅控制模块，用于若当前的资源服务器的质量浮动值表征的跌幅达到预设跌幅阈值，保持所述当前的资源服务器的负载权重值不变，并提高目标域名下除所述当前的资源服务器之外的其它资源服务器的负载权重值。
在一个实施例中，所述系统还包括：
调度策略下发单元，用于在针对目标域名下的各个所述资源服务器完成负载权重值调整后，将表征调整后的各个负载权重值的调度策略发送至DNS权威服务器，以使得所述DNS权威服务器按照所述调度策略表征的各个负载权重值，将指向所述目标域名的访问请求在各个所述资源服务器之间进行分配。
本申请还提供一种资源服务器的负载调节设备，所述设备包括处理器和存储器，所述存储器用于存储计算机程序，所述计算机程序被所述处理器执行时，可以实现上述的资源服务器的负载调节方法。
本申请还提供一种计算机存储介质，所述计算机存储介质用于存储计算机程序，所述计算机程序被处理器执行时，可以实现上述的资源服务器的负载调节方法。
所述存储器和计算机存储介质，可以包括用于存储信息的物理装置，通常是将信息数字化后再以利用电、磁或者光学等方法的媒体加以存储。本实施 例所述的存储器和计算机存储介质又可以包括：利用电能方式存储信息的装置，如RAM、ROM等；利用磁能方式存储信息的装置，如硬盘、软盘、磁带、磁芯存储器、磁泡存储器、U盘；利用光学方式存储信息的装置，如CD或DVD。当然，还有其他方式的存储器，例如量子存储器、石墨烯存储器等等。
在本实施例中，所述处理器可以按任何适当的方式实现。例如，所述处理器可以采取例如微处理器或处理器以及存储可由该(微)处理器执行的计算机可读程序代码(例如软件或固件)的计算机可读介质、逻辑门、开关、专用集成电路(Application Specific Integrated Circuit，ASIC)、可编程逻辑控制器和嵌入微控制器的形式等等。
由上可见，本申请提供的技术方案，针对记录的运行数据，可以进行多个检测周期的检测。在当前检测周期内，可以计算出资源服务器的历史特征值，该历史特征值可以用于表征资源服务器截至目前为止的服务性能。基于该历史特征值和当前检测周期内记录的运行数据，可以确定该资源服务器在当前检测周期内的质量浮动值。该质量浮动值可以作为该资源服务器的负载权重值的调整依据。后续，可以读取该资源服务器在上一个检测周期结束后的负载权重值，该负载权重值便可以是资源服务器当前采用的负载权重值。然后，可以根据上述计算出的质量浮动值，对该资源服务器当前采用的负载权重值进行动态调整，以使得调整后的负载权重值，与该资源服务器当前的处理性能相匹配。这样，当资源服务器当前的处理性能下降时，分配至该资源服务器的访问流量也会适当减少；而当资源服务器当前的处理性能提升时，分配至该资源服务器的访问流量也会适当增多。因此，通过动态调节资源服务器的负载权重值，能够提升系统的整体稳定性。
通过以上的实施例的描述，本领域的技术人员可以清楚地了解到各实施例可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件来实现。基于这样的理解，上述技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中，如ROM/RAM、磁碟、光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行各个实施例或者实施例的某些部分所述的方法。
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 (18)

  1. 一种资源服务器的负载调节方法，应用于调度系统中，包括：
    根据资源服务器的历史特征值，以及当前检测周期内记录的运行数据，确定所述资源服务器在所述当前检测周期内的质量浮动值；
    读取所述资源服务器当前采用的负载权重值，并基于确定的所述质量浮动值对所述资源服务器当前采用的负载权重值进行调整，并将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值。
  2. 根据权利要求1所述的方法，其中，所述方法还包括：
    根据监控探测方式和监控探测频率，为所述资源服务器生成监控任务，并在所述监控任务执行时，记录所述资源服务器对应的监控结果；
    根据通信地址，采集位于所述通信地址处的资源服务器的运行日志；
    将记录的所述监控结果和采集的运行日志作为所述资源服务器的运行数据。
  3. 根据权利要求2所述的方法，其中，所述方法还包括：
    从所述运行数据中，识别出携带无效状态码的运行数据，并将所述携带无效状态码的运行数据从记录的运行数据中剔除；
    从所述运行数据中，识别出表征所述资源服务器的资源不可用的运行数据，并将表征所述资源服务器的资源不可用的运行数据从记录的运行数据中剔除。
  4. 根据权利要求1所述的方法，其中，还包括以下方法：
    计算所述当前检测周期内记录的各条运行数据的特征值之和；
    读取上一个检测周期结束后所述资源服务器的历史特征值，并根据读取的所述历史特征值，和位于所述当前检测周期之前的历史运行数据的数据量，计算所述历史运行数据的特征值之和；
    根据所述当前检测周期内计算得到的特征值之和以及所述历史运行数据的特征值之和，计算所述当前检测周期内所述资源服务器的历史特征值。
  5. 根据权利要求4所述的方法，其中，计算所述当前检测周期内记录的各 条运行数据的特征值之和包括：
    累计所述当前检测周期内，各条运行数据的特征值，并将累计的结果作为所述当前检测周期内各条运行数据的特征值之和；
    其中，若当前运行数据的特征值大于预设阈值，将所述当前运行数据的特征值修改为所述预设阈值。
  6. 根据权利要求4所述的方法，其中，计算所述历史运行数据的特征值之和包括：
    将读取的所述历史特征值和所述历史运行数据的数据量的乘积，作为所述历史运行数据的特征值之和。
  7. 根据权利要求4所述的方法，其中，计算所述当前检测周期内所述资源服务器的历史特征值包括：
    计算所述当前检测周期内记录的运行数据和所述历史运行数据的总数据量；
    累计所述当前检测周期内计算得到的特征值之和与所述历史运行数据的特征值之和，并将累计的结果与所述总数据量的比值，作为所述当前检测周期内所述资源服务器的历史特征值。
  8. 根据权利要求1所述的方法，其中，确定所述资源服务器在所述当前检测周期内的质量浮动值包括：
    在所述当前检测周期内记录的运行数据中，识别出特征值大于所述历史特征值的第一运行数据，以及识别出特征值小于或者等于所述历史特征值的第二运行数据；
    针对所述第一运行数据和所述第二运行数据，分别计算质量浮动值，并将计算得到的质量浮动值之和，作为所述资源服务器在所述当前检测周期内的质量浮动值。
  9. 根据权利要求1所述的方法，其中，基于确定的所述质量浮动值对所述资源服务器当前采用的负载权重值进行调整包括：
    若确定的所述质量浮动值位于指定浮动区间内，保持所述资源服务器当前采用的负载权重值不变；
    若确定的所述质量浮动值大于所述指定浮动区间的上限值，按照预设调整幅值提高所述资源服务器当前采用的负载权重值；
    若确定的所述质量浮动值小于所述指定浮动区间的下限值，按照所述预设调整幅值减小所述资源服务器当前采用的负载权重值。
  10. 根据权利要求1或9所述的方法，其中，将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值包括：
    若调整后的负载权重值位于指定负载区间内，将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值；
    若调整后的负载权重值大于所述指定负载区间的上限值，将所述指定负载区间的上限值作为所述资源服务器在所述当前检测周期结束后的负载权重值；
    若调整后的负载权重值小于所述指定负载区间的下限值，将所述指定负载区间的下限值作为所述资源服务器在所述当前检测周期结束后的负载权重值。
  11. 根据权利要求1所述的方法，其中，基于确定的所述质量浮动值对所述资源服务器当前采用的负载权重值进行调整包括：
    若当前的资源服务器的质量浮动值表征的跌幅达到预设跌幅阈值，保持所述当前的资源服务器的负载权重值不变，并提高目标域名下除所述当前的资源服务器之外的其它资源服务器的负载权重值。
  12. 根据权利要求1所述的方法，其中，所述方法还包括：
    在针对目标域名下的各个所述资源服务器完成负载权重值调整后，将表征调整后的各个负载权重值的调度策略发送至DNS权威服务器，以使得所述DNS权威服务器按照所述调度策略表征的各个负载权重值，将指向所述目标域名的访问请求在各个所述资源服务器之间进行分配。
  13. 一种资源服务器的负载调节系统，包括：
    质量浮动值确定单元，用于根据资源服务器的历史特征值，以及当前检测 周期内记录的运行数据，确定所述资源服务器在所述当前检测周期内的质量浮动值；
    负载权重值调整单元，用于读取所述资源服务器当前采用的负载权重值，并基于确定的所述质量浮动值对所述资源服务器当前采用的负载权重值进行调整，并将调整后的负载权重值作为所述资源服务器在所述当前检测周期结束后的负载权重值。
  14. 根据权利要求13所述的系统，其中，所述系统还包括历史特征值计算单元，所述历史特征值计算单元包括：
    当前检测周期计算模块，用于计算所述当前检测周期内记录的各条运行数据的特征值之和；
    历史运行数据计算模块，用于读取上一个检测周期结束后所述资源服务器的历史特征值，并根据读取的所述历史特征值，和位于所述当前检测周期之前的历史运行数据的数据量，计算所述历史运行数据的特征值之和；
    历史特征值计算模块，用于根据所述当前检测周期内计算得到的特征值之和以及所述历史运行数据的特征值之和，计算所述当前检测周期内所述资源服务器的历史特征值。
  15. 根据权利要求13所述的系统，其中，所述质量浮动值确定单元包括：
    运行数据识别模块，用于在所述当前检测周期内记录的运行数据中，识别出特征值大于所述历史特征值的第一运行数据，以及识别出特征值小于或者等于所述历史特征值的第二运行数据；
    质量浮动值计算模块，用于针对所述第一运行数据和所述第二运行数据，分别计算质量浮动值，并将计算得到的质量浮动值之和，作为所述资源服务器在所述当前检测周期内的质量浮动值。
  16. 根据权利要求13所述的系统，其中，所述系统还包括：
    调度策略下发单元，用于在针对目标域名下的各个所述资源服务器完成负载权重值调整后，将表征调整后的各个负载权重值的调度策略发送至DNS权威服务器，以使得所述DNS权威服务器按照所述调度策略表征的各个负载权重值， 将指向所述目标域名的访问请求在各个所述资源服务器之间进行分配。
  17. 一种资源服务器的负载调节设备，包括处理器和存储器，所述存储器用于存储计算机程序，所述计算机程序被所述处理器执行时，实现如权利要求1至12中任一所述的方法。
  18. 一种计算机存储介质，所述计算机存储介质用于存储计算机程序，所述计算机程序被处理器执行时，实现如权利要求1至12中任一所述的方法。
PCT/CN2019/090320 2019-05-09 2019-06-06 一种资源服务器的负载调节方法及系统、设备 WO2020224024A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP19874757.8A EP3761609B1 (en) 2019-05-09 2019-06-06 Load adjustment method and device for resource server
US16/869,141 US11108695B2 (en) 2019-05-09 2020-05-07 Method, system and device for adjusting load of resource server

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910383394.6A CN110191159B (zh) 2019-05-09 2019-05-09 一种资源服务器的负载调节方法及系统、设备
CN201910383394.6 2019-05-09

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/869,141 Continuation US11108695B2 (en) 2019-05-09 2020-05-07 Method, system and device for adjusting load of resource server

Publications (1)

Publication Number Publication Date
WO2020224024A1 true WO2020224024A1 (zh) 2020-11-12

Family

ID=67716001

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/090320 WO2020224024A1 (zh) 2019-05-09 2019-06-06 一种资源服务器的负载调节方法及系统、设备

Country Status (4)

Country Link
US (1) US11108695B2 (zh)
EP (1) EP3761609B1 (zh)
CN (1) CN110191159B (zh)
WO (1) WO2020224024A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111010735A (zh) * 2019-11-20 2020-04-14 京信通信系统（中国）有限公司 资源配置方法、装置、基站和存储介质
CN110891087B (zh) * 2019-11-22 2022-10-25 深圳市网心科技有限公司 一种日志传输方法、装置及电子设备和存储介质
CN111158454A (zh) * 2019-12-20 2020-05-15 青岛海尔科技有限公司 一种实时嵌入式操作系统低功耗方法及装置
CN113868110B (zh) * 2021-11-30 2022-07-12 阿里云计算有限公司 一种企业数字中台服务健康度的评估方法和装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160088072A1 (en) * 2014-09-19 2016-03-24 Facebook, Inc. Balancing load across cache servers in a distributed data store
CN106998340A (zh) * 2016-01-22 2017-08-01 大唐移动通信设备有限公司 一种板卡资源的负载均衡方法及装置
CN107124472A (zh) * 2017-06-26 2017-09-01 杭州迪普科技股份有限公司 负载均衡方法及装置、计算机可读存储介质
CN107196869A (zh) * 2017-07-14 2017-09-22 北京知道创宇信息技术有限公司 基于主机实际负载的自适应负载均衡方法、设备和系统
CN107295048A (zh) * 2016-04-01 2017-10-24 百度在线网络技术（北京）有限公司 负载均衡方法及装置
CN109120715A (zh) * 2018-09-21 2019-01-01 华南理工大学 一种云环境下动态负载均衡方法

Family Cites Families (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6754706B1 (en) * 1999-12-16 2004-06-22 Speedera Networks, Inc. Scalable domain name system with persistence and load balancing
IL152642A0 (en) * 2000-05-05 2003-06-24 Nomadix Inc Network usage monitoring device and associated method
CN101141817B (zh) * 2007-10-12 2010-12-01 中兴通讯股份有限公司 根据网关负荷进行承载分担的调度方法、装置及通信系统
CN101753444B (zh) * 2009-12-31 2012-01-18 卓望数码技术（深圳）有限公司 一种负载均衡方法和负载均衡装置
CN103368864A (zh) * 2013-07-31 2013-10-23 北京华易互动科技有限公司 一种基于c/s架构的智能负载均衡方法
US9917781B2 (en) * 2014-06-05 2018-03-13 KEMP Technologies Inc. Methods for intelligent data traffic steering
US9979674B1 (en) * 2014-07-08 2018-05-22 Avi Networks Capacity-based server selection
US9774537B2 (en) * 2014-09-30 2017-09-26 Nicira, Inc. Dynamically adjusting load balancing
US9888063B2 (en) * 2014-12-10 2018-02-06 International Business Machines Corporation Combining application and data tiers on different platforms to create workload distribution recommendations
US10050927B2 (en) * 2015-01-27 2018-08-14 Mastercard International Incorporated Systems and methods for centralized domain name system administration
US10187457B1 (en) * 2015-12-23 2019-01-22 Github, Inc. Using read weight affinities to optimize the reading of block cache data
CN108009016B (zh) * 2016-10-31 2021-10-22 华为技术有限公司 一种资源负载均衡控制方法及集群调度器
US20200026868A1 (en) * 2018-07-23 2020-01-23 Wasabi Technologies, Inc. Methods and systems for storing data sequentially on a media

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160088072A1 (en) * 2014-09-19 2016-03-24 Facebook, Inc. Balancing load across cache servers in a distributed data store
CN106998340A (zh) * 2016-01-22 2017-08-01 大唐移动通信设备有限公司 一种板卡资源的负载均衡方法及装置
CN107295048A (zh) * 2016-04-01 2017-10-24 百度在线网络技术（北京）有限公司 负载均衡方法及装置
CN107124472A (zh) * 2017-06-26 2017-09-01 杭州迪普科技股份有限公司 负载均衡方法及装置、计算机可读存储介质
CN107196869A (zh) * 2017-07-14 2017-09-22 北京知道创宇信息技术有限公司 基于主机实际负载的自适应负载均衡方法、设备和系统
CN109120715A (zh) * 2018-09-21 2019-01-01 华南理工大学 一种云环境下动态负载均衡方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3761609A4 *

Also Published As

Publication number Publication date
EP3761609B1 (en) 2021-12-15
EP3761609A4 (en) 2021-01-13
US20200382423A1 (en) 2020-12-03
EP3761609A1 (en) 2021-01-06
CN110191159A (zh) 2019-08-30
CN110191159B (zh) 2021-05-25
US11108695B2 (en) 2021-08-31

Similar Documents

Publication Publication Date Title
WO2020224024A1 (zh) 一种资源服务器的负载调节方法及系统、设备
CN106933650B (zh) 云应用系统的负载管理方法及系统
US20200287794A1 (en) Intelligent autoscale of services
US7840517B2 (en) Performance evaluating apparatus, method, and computer-readable medium
US9866573B2 (en) Dynamic malicious application detection in storage systems
EP3525398A1 (en) Protection method and system for cdn client source station
US9864749B2 (en) Methods for provisioning workloads in a storage system using machine learning and devices thereof
US7908314B2 (en) Method for controlling a management computer
US9825869B2 (en) Traffic shaping based on request resource usage
WO2017000628A1 (zh) 一种云计算系统中的资源调度方法及装置
US9477618B2 (en) Information processing device, information processing system, storage medium storing program for controlling information processing device, and method for controlling information processing device
US20120221730A1 (en) Resource control system and resource control method
US20100046375A1 (en) Congestion Control Using Application Slowdown
US7817562B1 (en) Methods and systems for back end characterization using I/O sampling
CN112165436A (zh) 流量控制方法、装置及系统
CN109597800B (zh) 一种日志分发方法及装置
CN113190417A (zh) 微服务状态检测方法、模型的训练方法、设备及存储介质
CN111966918B (zh) 一种用于并发访问请求的限流方法、装置以及系统
CN116166443B (zh) 一种游戏任务系统的负载优化方法及系统
WO2023082765A1 (zh) 服务器状态控制方法、系统及存储介质
JP2017156888A (ja) 性能劣化要因推定プログラム、および性能劣化要因推定装置
CN109408302A (zh) 一种故障检测方法、装置及电子设备
JP2018022305A (ja) 境界値特定プログラム、境界値特定方法および境界値特定装置
CN107491264A (zh) 一种分布式系统中的数据写入方法及装置
US20180278645A1 (en) Software evaluation method and software evaluation device

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2019874757

Country of ref document: EP

Effective date: 20200428

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

Ref document number: 19874757

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE