Unlock instant, AI-driven research and patent intelligence for your innovation.

A method and device for dynamically adjusting task-level parameters of a distributed system architecture

A distributed system and dynamic adjustment technology, applied in the field of cloud computing, can solve the problem that parameter configuration cannot be adjusted according to user needs, and achieve the effect of online performance optimization.

Active Publication Date: 2019-04-26
SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] What the present invention aims to solve is the technical problem that the parameter configuration of each task cannot be dynamically adjusted according to user needs during the operation process

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A method and device for dynamically adjusting task-level parameters of a distributed system architecture
  • A method and device for dynamically adjusting task-level parameters of a distributed system architecture
  • A method and device for dynamically adjusting task-level parameters of a distributed system architecture

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0017] Embodiment 1, a method for dynamically adjusting task-level parameters of a distributed system architecture, such as figure 1 shown, including:

[0018] Step 1, create a configuration service thread when the job object is initialized, and this thread checks whether there is a new p group configuration every predetermined time;

[0019] Specifically, the technical solution of the present invention creates a separate configuration service thread, and sets every second or other predetermined time, such as N seconds, to monitor the update of the configuration file at the user's given location and whether the job is completed. Among them, p is a positive integer.

[0020] Step 2, if there are resources, allocate resources to them first;

[0021] Specifically, the technical solution of the present invention designs a resource re-application mechanism. Assuming that the user provides p groups (p, and the data structure is set as Config. The resource re-application mechanis...

Embodiment 2

[0024] Embodiment 2, a method for dynamically adjusting task-level parameters of a distributed system architecture, based on Embodiment 1, as follows Figures 2 to 3 shown, including:

[0025] More preferably, the step 1, if there is one, assign resources to it preferentially. Specifically, when the p group configuration parameters are updated, then notify the job assignment object, and the job assignment object includes a resource re-application mechanism, and the resource re-application The application mechanism resubmits to the resource scheduler a resource application with a priority higher than the task priority of the non-dynamic configuration according to the parameter value of the task resource allocation in the p group configuration, and the resource scheduler will give priority to the resource scheduler after receiving the application. p resource requests to allocate resources.

[0026] More preferably, when the step-p group of configurations is updated, it is assum...

Embodiment 3

[0042] Embodiment 3, a device for dynamically adjusting task-level parameters of a distributed system architecture, such as Figure 4 As shown, it includes: a configuration service module, which is used to create a configuration service thread when the job object is initialized, and the thread checks whether there is a new p-group configuration every predetermined time; a resource re-application module, which is used if there is a new p-group The configuration assigns resources to it first; the second configuration module is used for each task to complete the configuration of the second parameter when the task is initialized after retrieving its own configuration through its own id.

[0043] More preferably, the configuration service module, if there is a new p-group configuration, then preferentially allocate resources to it. Specifically, when the p-group configuration parameters are updated, then notify the job assignment object, and the job assignment object includes resour...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a method and a device for dynamically adjusting task-level parameters of a distributed system architecture, and relates to the technical field of cloud computing; it solves the technical problem that the parameter configuration of each task cannot be dynamically adjusted according to user needs during the operation process; The technical solution includes: Step 1, creating a configuration service thread when the job object is initialized, and the thread will check whether there is a new p group configuration every predetermined time; Step 2, if there is, assign resources to it first; Step 3, After each task retrieves its own configuration through its own id, the second parameter configuration is completed when the task is initialized.

Description

technical field [0001] The invention relates to the technical field of cloud computing, in particular to a method and device for dynamically adjusting task-level parameters of a distributed system architecture. Background technique [0002] The idea of ​​the MapReduce programming model comes from the functional programming language Lisp, which was proposed by Google in 2004 and first applied to large clusters. At present, Hadoop / MapReduc, as a distributed big data processing platform, is the most successful open source implementation of this programming model. Due to its good scalability and fault tolerance, it has been more and more widely used. [0003] Hadoop currently has as many as 190 configuration parameters, of which about 20 parameters have a significant impact on the performance of MapReduce jobs. Through these parameters, three aspects of performance tuning can be performed: CPU, memory and I / O optimization. [0004] These parameters can be divided into two cate...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): H04L12/24H04L29/08
CPCH04L41/08H04L67/34H04L67/60
Inventor 贝振东曾经纬喻之斌须成忠张慧玲
Owner SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI