The Standards on a Cloud Intelligence Service Framework and Protocol for Construction, Deployment,and Publishing of No-Code Scalable Web Software Platform

3.2.1. Application Management Module

The application management module of the platform MUST support applications developed by management tenants, including:¶

• Adding application¶
• Deleting application¶

3.2.2. End User Management Module

The end-user management module of the platform MUST support the management of end-user information for all applications of the tenant, including:¶

• End-user registration¶
• The end user changes the password¶
• End-user logout¶

3.2.3. Relationship Management Module between the End User and the Application

The relationship management module between the end user of the platform and the application MUST support the management of the use relationship between the end user and the application, including:¶

• Establishment of user relationship¶
• Elimination of the use relationship¶

3.2.4. Resource Management Module

The resource management module of the platform MUST include at least:¶

• Structured data resource management module¶

  • The structured data resource management module is used to manage the tenant's structured data in the cloud platform, including management of structured data table information, management of structured data table data resources. Each row of structured data resources will belong to a unique table, and will also belong to a unique user.¶

• File resource management module¶

  • The file resource management module is used to manage the file resources of the tenant in the cloud platform. Each file resource will belong to the only application and the only user.¶

3.2.5. Fine-grained Access Rights Management Module

The fine-grained access authority management module of the platform MUST support the management of end-user access rights to fine-grained resources within the tenant, including:¶

• Structured form row data resource authority management module¶

  • The structured form row data resource authority management module includes:¶

    • For each structured table, configure the application of the structured table;¶
    • For each structured table, configure the access rights of the end user of the application to which the structured table belongs;¶
    • For each structured table, configure the access rights of the end user of the application that does not use the structured table;¶
    • For each structured table, configure the access rights of the end user to the single-row data resources that do not belong to the structured table.¶

• Single file resource authority management module¶

  • The single file resource authority management module includes:¶

    • For each application, configure the access rights of the end user using the application to file resources within the application;¶
    • For each application, configure the access rights of the end user who does not use the application to the file resources within the application;¶
    • For each application, configure end-user access to files that are not part of the application.¶

3.2.6. Billing Management Module

Billing Management Module is based on "Pay as Use", a billing method in Cloud Computing, and supports fine-grained resource billing management, including:¶

• Billing Management Submodule¶

  • The billing management submodule includes:¶

    • Manage billing events such as billing type, billing name, service unit price. Tenants configure corresponding billing events for different services in the application instance. End users trigger the billing events when the service is running.¶
    • Configure the running status(enable/disable) of the billing mechanism.¶

• Expense Settlement Submodule¶

  • The expense settlement submodule includes:¶

    • Expense settlement on bills generated by end users using paid tenant cloud services.¶
    • End users recharge balance when the balance is insufficient in the settlement process.¶

• Bill Submodule¶

  • The bill submodule includes:¶

    • manage the bills generated by end users.¶
    • End users can check the service consumption through the bills.¶
    • Tenants summarize and cout the user bills in the tenant's domain by billing.¶

3.2.7. Customer Service Management Module

Customer Service Management Module implements online customer service based on Netty and supports the platform to configure multi-tenant customer system, including:¶

• Communication Protocol¶

  • The communication protocol includes:¶

    • The protocol is divided into message header and message body.¶

    • Message header includes magic number, protocol version, serialization algorithm, message type, status, reserved fields, data length and data type, totaling 23 bytes.¶

      • Magic number is used to prevent anyone from sending data to the server port at will.¶
      • Serialization algorithm includes JSON serialization, Hessian serialization and ProtoBuffer serialization.¶
      • Message type includes sending message, receiving message, enabling customer service, relinking, applying for manual customer service, ending manual customer service, and so on.¶
      • Status is used to mark whether the request is normal.¶
      • Reserved fields are used to support Protocol upgrade.¶
      • Data type includes text, emotion, files, and so on.¶
      • Message body contains the message content that is actually sent. The fixed field length and data length can be used as the basis for judging whether the message is complete.¶

• Implementation Mechanism¶

  • The implementation mechanism includes:¶

    • Using Master-slave Reactor Multithreading model provided by Netty is to ensure lots of connections and I/O operations performance in Concurrent scenarios.¶
    • Bring in tenant ID. The module supports mutli-tenant form of customer service mechanism. When users and customer service send and receive messages, the tenant ID can distinguish apply customer service used by different tenants to realize the multi-tenancy of customer service mechanism.¶

3.2.8. Ways of Accessing Cloud Storage Resources

The platform SHOULD be able to access cloud storage resources in the following ways:¶

• Through the POST, DELETE, PUT, GET requests initiated by REST API, the structured data resources specified by structured data URI are added, deleted, modified and queried.¶
• Through the POST, DELETE and GET requests initiated by REST API, the file resources specified by file URI are uploaded, deleted and downloaded.¶

3.2.9. Format of Describing the Structured Data URI

The platform adopts the access method of the resources of the cloud storage platform described in 3.2.8, and it SHOULD be able to describe the structured data URI in the following format:¶

• /{structure table name}[/{filtered fields + combined form}/{conditions for filtered fields}];¶

Where, the structure table name is equivalent to the table name to be operated in the SQL statement. The contents in the brackets ([ ]) are optional: "NONE", which means all resources in the collection are manipulated. "YES", it represents an operation on a particular resource in the collection, equivalent to the WHERE clause in SQL.¶

3.2.10. Contents of the Ways of Accessing Cloud Storage Resources

The platform adopts a cloud storage platform resource access method described in 3.2.8, which SHOULD include:¶

• Initiate PUT requests for structured data resources to modify the data resources specified by the URI, including: overwrite and modify the specified fields, incremental changes to specified fields.¶
• Initiate GET requests for structured data resources to query the data resources specified by the URI. Specific return formats can be set for the data resources that are queried, including: return by the specified field, return by the specified information page, and return by the specified field filtering.¶

3.2.11. Cloud Machine Learning Module

The Cloud Machine Learning Module supports online configuration and learning of data and machine learning module, which MUST include:¶

• Data Conversion Module¶

  • The Data Conversion Module is used to convert the data into the format suitable for machine learning. All data generate new storage which doesn't affect the original data. The Data Conversion Module MUST include:¶

    • Column filtering, delete the useless column in the original data;¶
    • Column renaming, change the representation of the column name in the original data;¶
    • One-hot encoding, convert the categorical value in the original data into One-hot vectors;¶
    • Derivative value, treat multiple columns in the original data as variables, and then use logic operations, mathematical symbols, and mathematical formulas to generate new column values;¶
    • Value change, change the original column value to a new one with condition matching such as empty, non-empty, interval matching, regular expression matching, exact matching, contain, does not contain, equal, greater than, less than, greater than or equal to and less than or equal to.¶
    • Data cleaning, delete useless column value in the original column with condition matching such as empty, non-empty, interval matching, regular expression matching, exact matching, contain, does not contain, equal, greater than, less than, greater than or equal to, and less than or equal to.¶

• Machine Learning Models Configuration Module¶

  • Users can configure machine learning model and precompiled model in this module. The configured machine learning model MUST support complete parameter configuration, this module MUST include:¶

    • Support for classical machine learning algorithms, including generalized linear model, naive Bayesian model, support vector machine model, decision tree and random forest model, K-nearest neighbor model, and K-means model;¶
    • Support for deep neural networks, including input layer, output layer, fully connected layer, convolutional layer, pooling layer, and transpose convolution layer;¶
    • Use the converted data to precompile the machine learning model to ensure that the model configuration is correct.¶

• Machine Learning Task Queue Module¶

  • Machine Learning Task Queue Module is used to manage and sequence the machine learning tasks initiated by users. This module MUST have a global task queue to manage the machine learning tasks. The tasks are inserted into the queue according to the time of initiation. For each machine learning task, persistence SHOULD be finished before inserting.¶

• Machine Learning Resource Management Module¶

  • Machine Learning Resource Management Module is used to manage and schedule the resource for machine learning and MUST include:¶

    • Computing resource, which MUST include CPU and the basic environment for machine learning tasks to run, and can include GPU;¶
    • Storage resource, which MUST include disk and database. The data of images required for machine learning is stored in the disk and the general data is stored in the database;¶
    • Computing resources and storage resources can be distributed among multiple machine nodes. The scheduler works based on node load and machine learning tasks. Computing resources and storage resources SHOULD set thresholds. When the resource utilization reaches the threshold, the task schedule is no longer accepted.¶

3.3.1. System Administrator View

• User Management¶

  • This view SHOULD be able to manage the user information of the Web software system, including the functions of adding users, removing users, adding user group information for users, and deleting user group information for users.¶

• User Group Management¶

  • This view SHOULD be able to manage the user group information of the Web software system, including adding user groups, deleting user groups, adding user information for user groups, and deleting user information for user groups.¶

• Model Management¶

  • This view SHOULD be able to manage the model information of the Web software system, including adding model information, deleting model information, adding model field information, deleting model field information, modifying model field information, and assigning different access rights to different user groups for the model.¶

• Machine Learning Task Management¶

  • This view SHOULD manage machine learning tasks initiated by users in the web software system, including canceling tasks exchanging the queue order among multiple tasks.¶

3.3.2. User View

• Data Management¶

  • This view SHOULD be able to manage the data table records of the Web software system, including adding data records, deleting data records, modifying data records, searching data records, and obtaining all data records.¶

• Data Statistics¶

  • This view SHOULD be able to make statistics on the data table records of the Web software system, including functions such as record statistics, maximum statistics, minimum statistics, and average statistics for data table integral fields.¶

• Import Excel and Other Format Tables¶

  • This view SHOULD be able to import data from Excel tables (sheets) and other format tables into the corresponding tables of the Web software system.¶

• Export Excel and Other Formats¶

  • This view SHOULD be able to export data from the table of the Web software system to Excel tables (sheets) and tables in other formats.¶

• Advanced Search¶

  • This view SHOULD be able to perform multiple logical combinatorial searches for multiple fields of the table to query the record information that satisfies the search criteria.¶

• Data Conversion¶

  • This view SHOULD support data conversion of the data table in the web software system, including column filtering, column renaming, One-hot encoding, derivative value, value change, and data cleaning.¶

• Machine Learning¶

  • This view SHOULD support configuration and pre-compilation of machine learning model in the web software system. After the model and data conversion configuration are well done, this view can initiate machine learning tasks.¶

3.4.1. System Requirements Table and Data Record Exchange Table

The table described in this method MUST include the system requirements table and the data record exchange table:¶

3.4.2. Steps of Reading System Requirements Table

In the workflow of the Web software generation principle of this method, it MUST include the following steps:¶

1. Using the traversal unit of system requirements table to read through each cell of the system requirements table.¶
2. Using the reading unit of system administrator information to read system administrator ID, system administrator password, system title and other information when traversing system administrator information.¶
3. Using the reading unit of user group information to read user group ID information when traversing user group information.¶
4. Using the reading unit of user information to read user ID, user name, user password, user group ID (the user group ID must be included in user group information) when traversing user information.¶
5. Using the reading unit of model information to read the information such as the reading and writing rights of the user group to which it belongs, the reading and writing rights of other users, the model name, the inputting ID, the group ID to which it belongs, the field type, the field name and whether it is allowed to be empty, when traversing the model information.¶

3.4.3. Steps of Analyzing System Requirements Table

In the workflow of Web software generation of this method, it MUST include the analysis and verification flow of the relevant information about the system requirements acquired by the Web software framework, including the following parts:¶

1. Data type and format validation for information related to system requirements obtained.¶
2. As for the user information part of the system requirements table, if the user's group item is not empty, the framework should judge whether the filled value is included in the user group information part of the system requirements table.¶

3.4.4. Contents of Modules for Users

The workflow of the Web software generation principle of this method MUST include the module for users to use in the cloud environment where the corresponding Web software framework is located, which must include the following contents:¶

1. The framework should use the API of dealing with system administrator information in the Read_Demand class to read the system administrator section in the software system requirement table,each cell of which will be read successively, and then register the software system using the super administrator privileges with its ID and password and the system name, and further create relevant database entities of the software system in the underlying database of the Web software framework. At the same time, such four collections should be create in the database as user group, user, schema, and data.¶
2. The framework should use the API of dealing with user group information in the Read_Demand class to read the user group information in the software system requirement table, each cell of which be read successfully, and generate json-formatted strings based on the group ID information obtained, and then insert them into the user group set of database entities created in 1.¶
3. The framework should use the API of dealing with user information in the Read_Demand class to read the user information in the software system requirement table, each cell of each row of which be read successfully, and generate json-formatted strings according to the user information obtained, and then insert them into the user set of database entities created in 1.¶
4. The framework should read the model information section of the system requirements table through the API used to process model information in the requirements table reading class (Read_Demand class), read through the filled content in the model information cells, generate json-formatted strings based on model information, and insert them into the schema set of database entities created in 1.¶
5. The framework should operate the user group data in the group set of database entities in the underlying database through the user group management class (Manage_Group) and return the results to the interface of the Web software framework. The Manage_Group class implements the GManage_Group interface defined in the Web software framework.¶
6. The framework should operate the user data in the user set of the database entity in the underlying database through the user management class (Manage_User) and return the results to the interface of the Web software framework. The Manage_User class implements the GManage_User interface defined in the Web software framework.¶
7. The framework should operate the model data in the schema set of database entities in the underlying database through the model management class (Manage_Schema) and return the results to the interface of the Web software framework. The Manage_Schema class implements the GManage_Schema interface defined in the Web software framework.¶
8. The framework should operate the data records in the data set of database entities in the underlying database through the data management class (Manage_Data) and return the results to the interface of the Web software framework. The Manage_Data class implements the GManage_Data interface defined in the Web software framework.¶

3.4.5. Steps of Principle of Constructing Web Software System

This method MUST satisfy the working principle of constructing Web software system automatically according to the system demand table. The principle includes the following steps:¶

1. The customer shall fill in the system requirements form described in 3.4.1.1 according to the requirements.¶
2. The Web software framework obtains the system requirements table uploaded by the customer in 1, and verifies whether the system requirements table conforms to the verification standard. If it does not conform to the verification standard, it prompts the user to have the wrong format, please re-upload it.¶
3. The Web software framework reads the system administrator information in the system requirements table.¶
4. Create a new system in the corresponding Web software framework according to the information read in 3, in which the administrator of the system is the system administrator in 3, and the system name is the system name in 3.¶
5. The Web software framework determines whether there is any part of user group information in the system requirements table. If not, it jumps to 8; if there is, it goes 6.¶
6. The Web software framework reads information about user groups in the system requirements table, including user group IDs.¶
7. According to the user group information read in 6, the user group is generated in the system created in 4, and the relevant information of the user group is the relevant information read in 6.¶
8. The Web software framework determines whether there is any relevant information about users in the system requirements table. If there is no information about users, it will jump to 11; if there is, it will go to 9.¶
9. The Web software framework reads relevant information of users in the system requirements table, including user ID, user name, user password, and user group.¶
10. The Web software framework generates users in the system created in 4 according to the user information read in 9, and the relevant information of the users is the information read in 9.¶
11. The Web software framework determines whether the relevant information of the model part exists in the system requirements table. If it does not, it will jump to 20, and if it does, it will go to 12.¶
12. The Web software framework reads the number of models in the system requirements table and assigns a current count of 0, then go to 13.¶
13. The Web software framework reads the relevant information of the model in the system requirements table, including the model name, model entry person, model group ID, group permissions, other user rights and other information.¶
14. According to the information related to the model read in 13, the model is generated in the system created in 4. The model name, model entry person, group ID of the model, group permission and other user rights are the information read in 13.¶
15. The Web software framework determines whether there is any information related to the structural fields in the model in the system requirements table. If there is no information, it will jump to 18; if there is, it will go to 16.¶
16. The Web software framework reads the relevant information of structural fields in the model in the system requirements table, including obtaining the name of structural fields of the model, field types, whether to allow null identification, default values and alternative values, etc.¶
17. The Web software framework creates model fields in the specified model according to the relevant information of structural fields in the model read in 16, and the field information is the information read in 16.¶
18. Current model number +1.¶
19. The Web software framework determines whether the current count is less than the number of models read in 12; if it is less than, it will jump back to 13; if it is not less than, it will go to 20.¶
20. The Web software framework generates the target Web software system.¶

3.4.6. Steps of Injecting Data Records into Web Software System

The workflow of the Web software generation principle of this method MUST include the workflow of automatically injecting data records into the Web software system according to the data record exchange table, including the following steps:¶

1. The customer shall fill in the data record exchange form as required.¶
2. The Web software framework gets the data record exchange table uploaded by the customer, and verifies whether the data record exchange form meets the requirements of the data record exchange form. If it does not meet the requirements, it prompts the user to have the wrong format, please re-upload it.¶
3. The Web software framework reads the model field information in the data record exchange table.¶
4. The Web software framework determines the required data exchange model according to the model field information read in 3.¶
5. The Web software framework reads the data record information in the data record exchange table line by line and injects it into the model determined in 4 successively.¶

3.4.7. Submission of System Requirements Table

The submission of the system requirements table SHALL include:¶

1. Online submission: according to the Web form format provided by the Web software generation framework, users fill in the system requirements information online and submit it to the Web software generation framework;¶
2. Offline submission: according to the offline form format template agreed by the Web software generation framework, users fill in the required software system demand information offline and submit it to the Web software generation framework by file upload.¶

3.4.8. Submission of Data Exchange Table

The submission of the data exchange table SHALL include:¶

1. Online submission: after the user constructs a new Web software system according to the Web software generation framework, the user enters data into the Web page through the newly generated Web software system, enters data information online and delivers it to the generated software system;¶
2. Offline submission: after the user constructs a new Web software system according to the Web software generation framework, the user enters the data record information offline and submits it to the generated software system through the data record exchange table template format provided by the newly generated Web software system by means of file upload.¶

3.5.1. System Deployment

This framework SHOULD support two ways to deploy a system:¶

• Supports registering users' own systems through cloud platform administrators to create system administrators belonging to users.¶
• Supports users to register their own system through cloud platform and register their own system administrator.¶

3.5.2. Model Deployment

This framework SHOULD enable system administrator users to enter their own software system and deploy the created models on the cloud platform, where the created models are configured as follows:¶

• Model name, input personnel, and user group read and write permission are necessary when the model is created.¶
• Field name, field type, whether the field is allowed to be empty, field default value, and field alternative value in the model are optional fields for creating the model.¶

3.5.3. User Deployment

This framework SHOULD support to create new users affiliated to the software system created by a system administrator user on the cloud platform, where the creation of users is configured with user ID, user nickname, and user password.¶

3.5.4. User Group Deployment

This framework SHOULD support to create new users affiliated to the software system created by a system administrator user on the cloud platform, where the creation of user groups is configured as follows:¶

• The user group ID is necessary.¶
• It is optional to select many users into the user group, but one user group must have one use at least.¶

3.6.1. Model Migration

This framework SHOULD support the migration of the specified model from the source system to the target software system, where the source container of the model includes at least one of the following:¶

1. Relational Database¶

   • SQLSERVER¶
   • MYSQL¶
   • ORACLE¶

2. Excel File. Here, the model migration may support two modes:¶

   • Migrating data from source to specify model structure¶
   • Migrating data from source to specify the model structure and migrating data within the model¶

3.6.2. User Migration

This framework SHOULD support the migration of users in the source system and their access rights to the target software system, where the user's source container includes at least one of the following:¶

1. Relational Database¶

   • SQLSERVER¶
   • MYSQL¶
   • ORACLE¶
2. Excel File¶

3.6.3. User Group Migration

This framework SHOULD support the migration of user groups in the source system and their access rights to the target software system, where the user group's source container includes at least one of the following:¶

1. Relational Database¶

   • SQLSERVER¶
   • MYSQL¶
   • ORACLE¶
2. Excel File¶

3.6.4. User Group Deployment

This framework SHOULD enable system administrator users to enter their own software systems and create user groups on the cloud platform, where the creation of user groups is configured as follows:¶

1. A user group ID created for identifying the user group is of necessity.¶
2. It is optional to select many users into a user group, but one user group must have one use at least.¶