CN110532192A - A kind of api interface performance optimized experimental facility - Google Patents
A kind of api interface performance optimized experimental facility Download PDFInfo
- Publication number
- CN110532192A CN110532192A CN201910840107.XA CN201910840107A CN110532192A CN 110532192 A CN110532192 A CN 110532192A CN 201910840107 A CN201910840107 A CN 201910840107A CN 110532192 A CN110532192 A CN 110532192A
- Authority
- CN
- China
- Prior art keywords
- experiment
- module
- service
- name
- experimental
- Prior art date
- Legal status (The legal status 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 status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3604—Software analysis for verifying properties of programs
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of api interface performance optimized experimental facilities, including experiment management module, which creates experiment for defining entire interface link level call relation in corresponding level, experiment is being created, the Experiment name of every group of distribution can be corresponded to;Deployment module is tested, which carries out packing mirror image for optimization branch code corresponding to above-mentioned Experiment name, and distributes starting container;Service registration discovery module, the module is for routing corresponding request by service name, Experiment name.Beneficial effect is: can carry out multiple performance tests simultaneously, shorten construction period;Multilayer performance test is independent mutually, does not influence the accuracy of data statistics;100% flow is all enjoyed in every layer of experiment, it is ensured that enough experimental datas;Quickly creation experiment;Every layer of each experimental group deployed environment is physical isolation, and reducing external other factors influences experimental result;Experiment real time monitoring second grade stops loss.
Description
Technical field
The present invention relates to field of software engineering, and in particular to a kind of api interface performance optimized experimental facility.
Background technique
Performance is always to measure an important indicator of software systems, directly affects the usage experience of user.With system
The increase of concurrent user's quantity of receiving, how under the hardware conditions such as existing bandwidth, cpu, memory, hard disk, from code layer
Rapid Optimum iteration existing system is a huge test on face.Test under line simulates ring on true line due to being difficult
Border, it is only wired on truthful data can just provide most accurate performance feedback, therefore tested on line just and become particularly significant.It is existing
Technology under, if the multiple technical optimization schemes of validation test simultaneously in experiment, there may be mutual exclusion between each technical solution,
And being difficult determination in outcome evaluation is which prioritization scheme really improves performance indicator.The physics of every kind of technical solution consumption
Resource is also different, influences experiment knot if carrying out test simply by flow cutting and will receive physical resource change outer factor
The authenticity of opinion.It can only test every time in practice and verify seldom Technological improvement plan, significantly extend the delivery time, increase
Unnecessary test job amount and the work of duplicate data statistic analysis.
Influence of the prior art using Monolayer Assay test different technologies prioritization scheme to performance indicator, multiple technical optimizations
Scheme need point more than the phase it is online, and create new experiment every time, require additional exploitation investment, therefore new in original exploitation
The reaction speed for increasing manpower and time cost on the basis of function, increasing delivery cycle, delaying product to market, drop
Low exploitation Optimized Iterative speed.Another scheme of the prior art is to be multiple portions by traffic partition while being in progress multiple
Experiment, but excessive detailed traffic flow be easy to cause the very few reliability for influencing test result assessment of data.
Summary of the invention
The object of the invention is that a kind of api interface performance optimized experimental facility is provided to solve the above-mentioned problems,
By the way that experimental technique prioritization scheme is divided into multiple subsets being independent of each other, it is abstracted as different calling experiment levels, flow point
Summarize with scheme, data statistics type, resource utilization statistics, and by different experimental groups, pass through appearance in deployment
Device technology carries out physical vlan isolation, docks each layer of calling in such a way that service registration is found, facilitates and realizes interface optimization
Scheme control experiment excludes the influence that multi-scheme mixes portion's bring physical resource, and can be by runnability and working as
When the physical resource that uses do cross validation, carry out more accurate performance evaluation, so as to by modification configuration quickly, it is same
The multiple interfaces for calling level of Shi Jianli, which optimize, tests, mutually indepedent between different levels, and each level enjoys 100%
Flow, to not only guarantee multiple experiments while progress, but also can guarantee the enough experimental datas of each level, be conducive to obtain
More accurate experiment conclusion reduces manpower and time cost to accelerate iteration speed.
To achieve the above object, the present invention provides following technical schemes:
The present invention provides a kind of api interface performance optimized experimental facility, including experiment management module, the module is whole for defining
A interface link level call relation, and experiment is created in corresponding level, experiment is being created, every group of distribution can be corresponded to
Experiment name;
Deployment module is tested, which carries out packing mirror image for optimization branch code corresponding to above-mentioned Experiment name, and
Distribution starting container;
Service registration discovery module, the module is for routing corresponding request by service name, Experiment name;
Statistical analysis module, the module are used to count the performance indicator of current interface and the physics money of current interface place container
Source information, and show eventually by index in experiment porch, facilitate the corresponding prioritization scheme of analysis whether effectively or for next time
Iteration, which is improved, provides data support.
It is called directly between relationship level and every level preferably, the experiment management module is used to management service
The corresponding Experiment name of each experimental group and its experiment parameter, while also needing definition management assignment of traffic mode strategy.
Preferably, experimental arms are mainly carried out packing mirror image by the experiment deployment module, and container platform into
Row deployment, while the experiment deployment module also includes a statistical fractals, the statistical fractals are for monitoring server physics
Data are sent to experiment porch in real time, and are associated with by corresponding Experiment name with corresponding experimental group on experiment porch
Summarize.
Preferably, the service registration discovery module is integrated in branch code, the service registration discovery module
It needs business item to integrate a SDK in insert experiment platform, includes empirical flow allocation strategy and a service in the SDK
Discovery module is registered, service registration discovery module provides two functions, and one can be by the service when being the starting after service arrangement
Corresponding service name, Experiment name are registered to registration center, and upstream called side is facilitated to call the service;Another function is
When upstream called side determines to test this assignment of traffic to some by traffic distribution strategy, service discovery function can lead to
Cross Experiment name, service name finds the IP of corresponding service, and initiates to request.
Preferably, the statistical analysis module is mainly to calculate each performance for testing each control group in layering to refer to
Mark, the statistical analysis module real-time experiment with computing indices offline by the way of log.
Preferably, the experimental provision further includes experiment monitoring module, the module is for finding that any experiment bring is different
Often, and immediately alarm, under fall unreasonable experiment, thus make experiment management module backstage configuration at once stop loss.
The utility model has the advantages that the 1, present invention can carry out multiple performance tests simultaneously, shorten construction period;
2, multilayer performance test is independent mutually, does not influence the accuracy of data statistics;
3,100% flow is all enjoyed in every layer of experiment, it is ensured that enough experimental datas;
4, quickly creation experiment;
5, every layer of each experimental group deployed environment is physical isolation, and reducing external other factors influences experimental result;
6, experiment real time monitoring second grade stops loss.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention for those of ordinary skill in the art without creative efforts, can be with
It obtains other drawings based on these drawings.
Fig. 1 is flow chart of the invention;
Fig. 2 is the data structure diagram of experiment management module of the invention;
Fig. 3 is the data statistics journal format of statistical analysis module of the invention;
Fig. 4 is the AABB homogeneity experimental data table that the present invention uses.
The reference numerals are as follows:
1, experiment management module;2, deployment module is tested;3, service registration discovery module;4, statistical analysis module.
Specific embodiment
To make the object, technical solutions and advantages of the present invention clearer, technical solution of the present invention will be carried out below
Detailed description.Obviously, described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.Base
Embodiment in the present invention, those of ordinary skill in the art are obtained all without making creative work
Other embodiment belongs to the range that the present invention is protected.
Shown in Figure 1, the present invention provides a kind of api interface performance optimized experimental facilities, including experiment management module
1, which creates experiment for defining entire interface link level call relation in corresponding level, is creating
Experiment can correspond to the Experiment name of every group of distribution;
Deployment module 2 is tested, which carries out packing mirror image for optimization branch code corresponding to above-mentioned Experiment name, and
Distribution starting container;
Service registration discovery module 3, the module is for routing corresponding request by service name, Experiment name;
Statistical analysis module 4, the module are used to count the performance indicator of current interface and the physics of current interface place container
Whether effectively or be next resource information, and showing eventually by index in experiment porch facilitates the corresponding prioritization scheme of analysis
Secondary iteration, which is improved, provides data support.
It is called directly between relationship level and every level preferably, the experiment management module 1 is used to management service
The corresponding Experiment name of each experimental group and its experiment parameter, while definition management assignment of traffic mode strategy is also needed, for clothes
Business developer needs to integrate the SDK of an experiment porch in each service, and default experiment porch can draw 100% flow
It is divided into 1000 Traffic Domains, creates different experimental groups when creating experiment, and to specify different experiments group corresponding
Traffic Domain is used for subsequent assignment of traffic.The member needed in one experiment is known as: experiment unique identification (layer id), experiment
Traffic Domain range of distribution, the strategy etc. of assignment of traffic, these elements can easily be added in the page, be then stored in configuration
Management system facilitates unified obtain.One of the characteristics of program is exactly that can guarantee that each calling layer is enjoyed absolutely
Flow, in this scenario, it to be stochastic-flow side formula, user side mode, orientation flow group respectively that there are many traffic distribution strategies
Three kinds of modes.Whole flow is divided into 1000 parts, all methods of salary distribution are all to first pass through certain rule to calculate one
Then the experiment domain that hash % 1000, final result and experiment porch user fill in is compared and obtains by hash value
The domain that the request should distribute.It is mutually indepedent between each level in order to guarantee to test, each layer of experiment when calculating hash value
Can recalculate, and layer id is added, guarantee that each laminar flow amount can break up and redistribute, thus avoid layer with
Layer directly influences each other.For the method for salary distribution hash value of stochastic-flow side directly can be used hash (random,
Layer id) correlation function, hash (user id, layer id) is used for user side, for orienting flow group
Mode is analyzed by stereotactic conditions first, can enter this layer experiment when meeting stereotactic conditions;After experiment creation,
SDK can periodically with experiment porch synchrodata, data structure such as Fig. 2, respectively be experiment call lower service service name and
Empirical flow allocation strategy parameter book and the corresponding Experiment name of each grouping, experiment parameter.
Experimental arms are mainly carried out packing mirror image by the experiment deployment module 2, and are disposed in container platform, together
It also includes a statistical fractals that Shi Suoshu, which tests deployment module 2, which is used for server physics monitoring data is real-time
Be sent to experiment porch, and summarized on experiment porch by the association of corresponding Experiment name and corresponding experimental group.
The service registration discovery module 3 is integrated in branch code, and the service registration discovery module 3 is real in access
It needs business item to integrate a SDK when testing platform, includes empirical flow allocation strategy and a service registration discovery in the SDK
Module 3, service registration discovery module 3 provide two functions, and one can be corresponding by the service when being the starting after service arrangement
Service name, Experiment name are registered to registration center, and upstream called side is facilitated to call the service;Another function is to work as upstream
When called side determines to test this assignment of traffic to some by traffic distribution strategy, service discovery function can pass through experiment
Title, service name find the IP of corresponding service, and initiate to request.The main mesh of the module is integrated by experiment porch
Be solve creation experiment, empirical flow distribution be fully automated, without think by modification code realize calling ratio tune
Whole, service registration discovery is by the regular uploaded state of heartbeat to registration center, and for calling, service side can verify heartbeat every time
Time determines whether node survives, in order to guarantee the High Availabitity of business on line, if it find that experimental group is problematic or configures
Experimental group can default the service node for calling control group when not finding corresponding service node, and report experiment porch,
Experiment porch can indicate that current experiment is invalid state.
The statistical analysis module 4 is mainly the performance indicator for calculating each control group in each experiment layering, the system
Count the real-time experiment with computing indices offline by the way of log of analysis module 4.It unites particular by the SDK of experiment porch
Current performance indexes is counted, and all experiment layer ID, the Traffic Domain ID of present flow rate are finally write into log
In;It is to be grouped first to layer id that specific offline logs processing, which calculates, is then continued through to the data after grouping
Traffic Domain ID is aggregated into some experimental group of specific current experiment layer in each layer, finally shows on experiment porch each
Each group of performance control is tested, specific journal format is as shown in Figure 3.
The experiment porch scientific and reasonable for one, especially under multilayer experimental situation, the principle of multilayer performance test is
It is independent of each other between the experiment of different layers, this is optimal state, but may test setting for itself in practice
It counts and unreasonable, therefore the experimental provision further includes experiment monitoring module, experiment monitoring module is for finding that any experiment is brought
Exception, and alarm immediately, under fall unreasonable experiment, so that the configuration on 1 backstage of experiment management module be made to stop loss at once.Such as figure
Shown in 4, the present invention is by the way of the experiment of AABB homogeneity come the reliability of confirmatory experiment result.So-called AABB homogeneity experiment
It exactly needs to divide four empirical flow domains for an experiment, is that two groups of the same set of strategies of AA are tactful as BB respectively
Two groups of collection, such as the performance of two kinds of algorithms of AB is compared, such as A algorithm distribution 5%, B algorithm distribution 5%, then just needing
The A algorithm control group of creation two 5% and the control group of two 5% B algorithms, verifying need to compare AA, BB respectively before compareing AB
Whether two groups of performance indicator consistent, if consistent or illustrate empirical flow distribution in confidence interval rationally, not by
Pollution, the performance indicator that can just compare AB are finally drawn a conclusion.
The above description is merely a specific embodiment, but scope of protection of the present invention is not limited thereto, any
Those familiar with the art in the technical scope disclosed by the present invention, can easily think of the change or the replacement, and should all contain
Lid is within protection scope of the present invention.Therefore, protection scope of the present invention should be based on the protection scope of the described claims.
Claims (6)
1. a kind of api interface performance optimized experimental facility, it is characterised in that: including
Experiment management module, the module are created in corresponding level for defining entire interface link level call relation
Experiment is built, experiment is being created, the Experiment name of every group of distribution can be corresponded to;
Deployment module is tested, which carries out packing mirror image for optimization branch code corresponding to above-mentioned Experiment name, and
Distribution starting container;
Service registration discovery module, the module is for routing corresponding request by service name, Experiment name;
Statistical analysis module, the module are used to count the performance indicator of current interface and the physics money of current interface place container
Source information, and show eventually by index in experiment porch, facilitate the corresponding prioritization scheme of analysis whether effectively or for next time
Iteration, which is improved, provides data support.
2. a kind of api interface performance optimized experimental facility according to claim 1, it is characterised in that: the experiment management mould
Block is used to management service and calls directly the corresponding Experiment name of each experimental group between relationship level and every level and its experiment
Parameter, while also needing definition management assignment of traffic mode strategy.
3. a kind of api interface performance optimized experimental facility according to claim 1, it is characterised in that: mould is disposed in the experiment
Experimental arms are mainly carried out packing mirror image by block, and are disposed in container platform, while the experiment deployment module is also wrapped
Containing a statistical fractals, which is used to send experiment porch in real time for server physics monitoring data, and
Summarized by corresponding Experiment name with the association of corresponding experimental group on experiment porch.
4. a kind of api interface performance optimized experimental facility according to claim 1, it is characterised in that: the service registration hair
Existing module is integrated in branch code, and the service registration discovery module needs business item integrated in insert experiment platform
It is mentioned in one SDK, the SDK comprising empirical flow allocation strategy and a service registration discovery module, service registration discovery module
For two functions, one is that the corresponding service name of the service, Experiment name can be registered to note when starting after service arrangement
Volume center, facilitates upstream called side to call the service;Another function is when upstream called side is determined by traffic distribution strategy
When this assignment of traffic being tested to some surely, service discovery function can be found corresponding by Experiment name, service name
The IP of service, and initiate to request.
5. a kind of api interface performance optimized experimental facility according to claim 1, it is characterised in that: the statistical analysis mould
Block is mainly the performance indicator for calculating each control group in each experiment layering, and the statistical analysis module is by the way of log
The offline indices of experiment with computing in real time.
6. a kind of api interface performance optimized experimental facility according to claim 4, it is characterised in that: the experimental provision also wraps
Include experiment monitoring module, which alarms immediately for finding any experiment bring exception, under fall unreasonable experiment,
To make the configuration on experiment management module backstage stop loss at once.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910840107.XA CN110532192A (en) | 2019-09-06 | 2019-09-06 | A kind of api interface performance optimized experimental facility |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910840107.XA CN110532192A (en) | 2019-09-06 | 2019-09-06 | A kind of api interface performance optimized experimental facility |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110532192A true CN110532192A (en) | 2019-12-03 |
Family
ID=68667265
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910840107.XA Pending CN110532192A (en) | 2019-09-06 | 2019-09-06 | A kind of api interface performance optimized experimental facility |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110532192A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120239444A1 (en) * | 2011-03-15 | 2012-09-20 | Accenture Global Services Limited | Mvt optimization of business process modeling and management |
CN106341290A (en) * | 2016-08-31 | 2017-01-18 | 北京城市网邻信息技术有限公司 | Flow distribution server |
CN110008131A (en) * | 2019-04-12 | 2019-07-12 | 重庆天蓬网络有限公司 | A kind of method and device for the region AB experiment management realized based on algorithm |
CN110149396A (en) * | 2019-05-20 | 2019-08-20 | 华南理工大学 | A kind of platform of internet of things construction method based on micro services framework |
CN110166285A (en) * | 2019-04-28 | 2019-08-23 | 北京航空航天大学 | A kind of network security experiment porch building method based on Docker |
-
2019
- 2019-09-06 CN CN201910840107.XA patent/CN110532192A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20120239444A1 (en) * | 2011-03-15 | 2012-09-20 | Accenture Global Services Limited | Mvt optimization of business process modeling and management |
CN106341290A (en) * | 2016-08-31 | 2017-01-18 | 北京城市网邻信息技术有限公司 | Flow distribution server |
CN110008131A (en) * | 2019-04-12 | 2019-07-12 | 重庆天蓬网络有限公司 | A kind of method and device for the region AB experiment management realized based on algorithm |
CN110166285A (en) * | 2019-04-28 | 2019-08-23 | 北京航空航天大学 | A kind of network security experiment porch building method based on Docker |
CN110149396A (en) * | 2019-05-20 | 2019-08-20 | 华南理工大学 | A kind of platform of internet of things construction method based on micro services framework |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110166282B (en) | Resource allocation method, device, computer equipment and storage medium | |
CN111639363B (en) | Data analysis method based on block chain and edge computing server | |
CN104092756B (en) | A kind of resource dynamic distributing method of the cloud storage system based on DHT mechanism | |
US20150248461A1 (en) | Streaming query deployment optimization | |
CN107786616A (en) | Main frame intelligent monitor system based on high in the clouds | |
CN103176849B (en) | A kind of dispositions method of the cluster virtual machine based on resource classification | |
CN104504103B (en) | A kind of track of vehicle point insertion performance optimization method and system, information acquisition device, database model | |
CN108092813A (en) | Data center's total management system server hardware Governance framework and implementation method | |
CN106407830A (en) | Detection method and device of cloud-based database | |
WO2023098374A1 (en) | Network resource deployment method and apparatus, and electronic device and storage medium | |
CN105827678B (en) | Communication means and node under a kind of framework based on High Availabitity | |
Smolka et al. | Evaluation of fog application placement algorithms: a survey | |
CN105468619B (en) | Resource allocation methods and device for database connection pool | |
CN108282501A (en) | A kind of Cloud Server resource information synchronous method, device and system | |
CN109828790A (en) | A kind of data processing method and system based on Shen prestige isomery many-core processor | |
CN110442602A (en) | Data query method, apparatus, server and storage medium | |
CN109726004A (en) | A kind of data processing method and device | |
CN109033173A (en) | It is a kind of for generating the data processing method and device of multidimensional index data | |
CN112000657A (en) | Data management method, device, server and storage medium | |
Fan et al. | Node essentiality assessment and distributed collaborative virtual network embedding in datacenters | |
CN117744129B (en) | Intelligent operation and maintenance method and system based on CIM | |
CN116389486B (en) | Method and system for realizing operation analysis of multiple cloud resources | |
CN108287876A (en) | Support the power quality data service quality detection method and device of multiple format | |
CN110532192A (en) | A kind of api interface performance optimized experimental facility | |
Raza et al. | Empowering microservices: A deep dive into intelligent application component placement for optimal response time |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191203 |