Application development method and development platform

Abstract

The invention provides an application development method and a development platform. All primitives provided by all distributed computing frameworks are encapsulated, a unified data processing interface is provided, input and output data structures are unified, a unified input and output interface is provided, and an application is decoupled from all the distributed computing frameworks; when performing application development, a user only need to design configuration parameters, decompose an application development task into multiple data processing steps and determine the sequence of all the data processing steps; the development platform determines data processing logic realization classes corresponding to all the data processing steps through the configuration parameters, a preset data processing interface and a preset input and output interface are called, all the data processing logic realization classes are controlled to inherit the preset data processing interface, and the application development task can be completed after to-be-processed data is processed. Based on the application development method and the development platform, the study and development cost can be effectively lowered for developers.

Description

technical field [0001] The present invention relates to the field of computer computing, and more specifically, to an application program development method and a development platform. Background technique [0002] Traditional distributed application development usually includes the following steps: read the input data source to be processed; convert the input data into a custom data structure; write application logic, call the distributed computing framework interface, perform data processing and calculation, Different computing frameworks will provide their own processing interfaces; output calculation results; write main functions and configure various related configurations used by distributed jobs. [0003] At present, there are mainly open source distributed computing frameworks such as Hadoop, spark, and storm, which realize various distributed batch and streaming real-time computing scenarios. However, when developing applications directly based on these open source...

AI Technical Summary

Problems solved by technology

However, when developing applications directly based on these open source frameworks, developers need to understand various underlying framework technologies and framework details such as Hadoop, spark, and storm, and the learning cost is high

The application program is highly coupled with the underlying framework. If you need to add a new computing scenario to the existing open source distributed computing framework, or support a new business component, you need to modify a large number of application codes. The scalability of the computing framework is poor and high maintenance cost

[0004] Therefore, based on the existing open source distributed computing framework for application development, the development cost and maintenance cost are high

Images