System adaptive current limiting method based on server hardware and software indicators

Abstract

The present invention provides a system adaptive current limiting method based on server software and hardware indicators, comprising the following steps: 1) the server receives user data and monitors the status of the server in real time; executes step 2; 2) when the server reaches a preset condition, triggers the current limiting strategy, and executes step 3; 3) reduces the data received by the server. The present invention combines monitoring indicators of several dimensions such as Load, CPU usage, overall average RT, entry QPS, and Jmeter thread number, and through an adaptive flow control strategy, balances the system's entry flow and system load, allowing the system to run at the maximum throughput as much as possible while ensuring the overall stability of the system.

Description

Technical Field

[0001] The invention relates to a server processing method, and in particular to a system adaptive current limiting method based on server software and hardware indicators. Background Art

[0002] With the development of science and technology, Internet applications are becoming more and more common, and the technical requirements for network servers are becoming higher and higher. For a large server, its interface will be frequently called by the outside world. If no access flow control is added, the interface call frequency will be too high, affecting the stability of the system.

[0003] A server, also known as a server, is a device that provides computing services. Since servers need to respond to service requests and perform processing, they should generally have the ability to undertake and guarantee services. The composition of a server includes a processor, hard disk, memory, system bus, etc., which is similar to the general computer architecture. However, since it needs to provide highly reliable services, it has higher requirements in terms of processing power, stability, reliability, security, scalability, manageability, etc.

[0004] In the server system, the backend server allocates resources such as CPU and memory each time it processes a request. In the case of high concurrency or when the system is attacked by traffic, the CPU usage and memory usage of the backend server will remain high, thus posing a risk of crash.

[0005] Therefore, it is necessary to improve the existing technology. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a system adaptive current limiting method based on server software and hardware indicators.

[0007] In order to solve the above technical problems, the present invention provides a system adaptive current limiting method based on server software and hardware indicators, comprising the following steps:

[0008] 1) The server receives user data and monitors the status of the server in real time; execute step 2;

[0009] 2) When the server reaches the preset conditions, the current limiting strategy is triggered and step 3 is executed;

[0010] 3) Reduce the data received by the server.

[0011] As an improvement to the system adaptive current limiting method based on server hardware and software indicators of the present invention:

[0012] In step 1:

[0013] When the server receives user data, it performs feature marking on each user data. For example, if the parameters in the CPU of the user data, the read and write speed of the IO interface, the memory size parameters, etc. exceed the set value, they are marked with a label with a larger corresponding parameter. One user data may include several labels, so that the parameters of a certain type of user data can be quickly searched and limited.

[0014] The setting value is: Use twice the average value.

[0015] As a further improvement to the system adaptive current limiting method based on server hardware and software indicators of the present invention:

[0016] The preset conditions in step 2 include: a parameter in the CPU of the server reaches a parameter threshold, or the read and write of the IO interface of the server reaches a read and write threshold, or the memory size of the server reaches a memory threshold;

[0017] As a further improvement to the system adaptive current limiting method based on server hardware and software indicators of the present invention:

[0018] When the server does not trigger the current limiting strategy, only less than 1% of the data requests have a response time rt that exceeds the response threshold;

[0019] The response threshold setting method is:

[0020] rt threshold = minRT * number of cores * coefficient

[0021] minRT is the minimum response time rt in the server's historical data; the number of cores is the total number of cores in the server's CPU; the coefficient can be set manually.

[0022] As a further improvement to the system adaptive current limiting method based on server hardware and software indicators of the present invention:

[0023] In step 3: the server reduces the received data according to the region, user type, and user tag.

[0024] The flow control is based on the region: provinces, cities, and districts represent each region. When limiting the flow, the flow of user data in a specific province, city, or district can be limited, or the flow of user data in the province, city, or district that receives the most data can be limited;

[0025] The flow is limited according to the user type: user types include member users and ordinary users. When limiting the flow, the user data of ordinary users can be limited.

[0026] Limiting the flow based on user tags: Based on the parameters that meet the preset conditions, select user data with tags with larger corresponding parameters for flow limiting.

[0027] As a further improvement to the system adaptive current limiting method based on server hardware and software indicators of the present invention:

[0028] The current limiting methods include directly rejecting new user data, or putting new user data in a pending state and re-reading it after the current limiting strategy ends.

[0029] The technical advantages of the system adaptive current limiting method based on server hardware and software indicators of the present invention are:

[0030] The present invention combines monitoring indicators of several dimensions, such as Load, CPU usage, overall average RT, inlet QPS, and Jmeter thread number, and uses an adaptive flow control strategy to achieve a balance between the system's inlet traffic and the system's load, so that the system can run at the maximum throughput as much as possible while ensuring the overall stability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The specific implementation modes of the present invention are further described in detail below with reference to the accompanying drawings.

[0032] Figure 1 It is a flow chart of the system adaptive current limiting method based on server software and hardware indicators of the present invention. DETAILED DESCRIPTION

[0033] The present invention is further described below in conjunction with specific embodiments, but the protection scope of the present invention is not limited thereto.

[0034] Embodiment 1: System adaptive current limiting method based on server hardware and software indicators, such as Figure 1 As shown, the following steps are included:

[0035] 1) The server receives user data and monitors the status of the server in real time; execute step 2;

[0036] When the server receives user data, it performs feature marking on each user data. For example, if the parameters in the CPU of the user data, the read and write speed of the IO interface, the memory size parameters, etc. exceed the set value (for example, using twice the average value), it will be marked with a label with a larger corresponding parameter. One user data may include several labels, so that the parameters of a certain type of user data can be quickly searched and limited.

[0037] During R&D, the online API will be subjected to a stress test (such as jmeter, etc.) to evaluate the API type and define it based on the stress test results for classification.

[0038] 2) When the server reaches the preset conditions, the current limiting strategy is triggered and step 3 is executed;

[0039] The preset conditions include: when a parameter in the CPU of the server reaches a parameter threshold, or the read and write of the IO interface of the server reaches a read and write threshold, or the memory size of the server reaches a memory threshold, the current limiting strategy is executed, and step 3 is executed;

[0040] The server's CPU parameters include CPU load, CPU comprehensive utilization, and CPU jumps.

[0041] The query rate per second QSP of the server has a limit of 1.5w. If it exceeds the limit, the service response time rt will be too long. It is necessary to ensure that only less than 1% of the data requests exceed the response threshold in the response time rt. The response threshold can be set to 3s.

[0042] The response threshold setting method is:

[0043] rt threshold = minRT * number of cores * coefficient

[0044] minRT is the minimum value of the response time rt in the server's historical data; the number of cores is the total number of cores of the server's CPU; the coefficient can be set manually, for example, 2.5. By changing the size of the coefficient, the size of the response time rt can be changed.

[0045] The corresponding time rt is positively correlated with the query rate per second QSP. The larger the corresponding time rt is, the larger the query rate per second QSP is. It is necessary to ensure that the query rate QSP is within a safe range. The preset condition is set accordingly. If the threshold of the preset condition is not reached, the query rate QSP is within a safe range. If the threshold of the preset condition is reached, current limiting is required.

[0046] 3) Reduce the user data received by the server until the server does not meet the preset conditions, that is, the parameters in the server's CPU are lower than the parameter threshold, the read and write of the server's IO interface are lower than the read and write threshold, and the memory size of the server is lower than the memory threshold.

[0047] The server can reduce the received data based on region, user type, user tag, etc.

[0048] Based on regional flow control: provinces, cities, districts, etc. are used to represent various regions. When limiting flow, the flow of user data in a specific province, city, or district can be limited, or the flow of user data in the province, city, or district that receives the most data can be limited;

[0049] The flow is limited according to the user type: user types include member users, ordinary users, etc. When limiting the flow, the user data of ordinary users can be limited.

[0050] Limiting the flow based on user tags: Based on the parameters that meet the preset conditions, select user data with tags with larger corresponding parameters for flow limiting.

[0051] The current limiting methods include directly rejecting new user data, or putting new user data in a pending state and re-reading it after the current limiting strategy ends.

[0052] Finally, it should be noted that the above examples are only some specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments, and there are many variations. All variations that can be directly derived or associated with the content disclosed by a person skilled in the art should be considered as the protection scope of the present invention.

Claims

1. A system adaptive current limiting method based on server software and hardware indicators, characterized by: The following steps are involved: 1) The server receives user data and monitors the status of the server in real time; execute step 2; 2) When the server reaches the preset conditions, the current limiting strategy is triggered and step 3 is executed; 3) Reduce the data received by the server; In step 1: When the server receives user data, it will mark each user data with features. If the parameters in the CPU, the read and write speed of the IO interface, and the memory size of the user data exceed the set value, the corresponding parameter will be marked with a larger tag. One user data may include several tags, so that the parameters of a certain type of user data can be quickly searched and limited. The settings are: use twice the average value; The preset conditions in step 2 include: a parameter in the CPU of the server reaches a parameter threshold, or the read and write of the IO interface of the server reaches a read and write threshold, or the memory size of the server reaches a memory threshold; In step 3: the server reduces the received data and can reduce it according to the region, user type, and user tag; The flow control is based on the region: provinces, cities, and districts represent each region. When limiting the flow, the flow of user data in a specific province, city, or district can be limited, or the flow of user data in the province, city, or district that receives the most data can be limited; The flow is limited according to the user type: the user type includes member users and ordinary users. When limiting the flow, the user data of ordinary users can be limited; Limiting traffic based on user tags: Based on the parameters that meet the preset conditions, select user data with tags with larger corresponding parameters for limiting traffic; When the server does not trigger the current limiting strategy, only less than 1% of the data requests have a response time rt that exceeds the response threshold; The response threshold setting method is: rt threshold = minRT * number of cores * coefficient minRT is the minimum response time rt in the server's historical data; the number of cores is the total number of cores in the server's CPU; the coefficient is set manually.

2. The system adaptive current limiting method based on server software and hardware indicators according to claim 1 is characterized in that: The current limiting methods include directly rejecting new user data, or putting new user data in a pending state and re-reading it after the current limiting strategy ends.

Images