Distributed Processing System for Application Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In distributed processing systems, application migration is prolonged due to inefficient server selection based on load states, incomplete process checking, and sequential request processing, leading to increased downtime.
Innovation Solution
A distributed processing system with a management device and application servers that include an adapter unit, processor unit, dispatcher unit, and statistical information storage unit, which manages migration status and evaluates performance to select optimal servers for process distribution and migration, reducing downtime by ensuring all processes are completed before migration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If servers are selected for migration based on load states such as CPU usage rates, then server selection is simplified, but servers with short migration periods are not guaranteed to be selected, increasing the degeneracy period of time
Solution Approach 1:
The system performs preliminary actions by selecting servers for migration before actual migration occurs. The management device identifies candidate servers and enters them into a prior blockage state in advance, allowing the system to prepare migration targets proactively rather than reactively, thus reducing the overall degeneracy period while maintaining simple selection criteria based on load states
2Ease of operation
If migration timing is decided based on transition of CPU or memory usage rate, then process completion checking is simplified, but useless time is wasted and the degeneracy period of time is increased
Solution Approach 1:
The system implements feedback mechanisms where the management device continuously monitors the actual completion status of processes on selected servers. This feedback loop allows the system to accurately determine when migration timing is appropriate based on real process completion data rather than relying on indirect CPU or memory usage rate transitions, eliminating useless waiting time while maintaining simple monitoring operations
3Reliability
If all processes for requests received until prior blockage are completed before migration, then process completion is ensured, but the system must wait until all processes are completed, increasing migration time
Solution Approach 1:
The system segments the process completion verification into two independent parts: (1) verification that all processes received before prior blockage are completed, and (2) verification that no new processes are accepted during migration. This segmentation allows the system to ensure process completion reliability while reducing migration time by not requiring completion of all possible processes, only those that would have been received before the blockage state was established
Data Source
AI summary
Provided is a distributed processing system in which migration evaluation is performed based on performance information, statistical information of each process of an application, and the number of non-completed processes of each process of an application for an application server in which an old application is being operated, a migration target server group is decided so that migration is completed in a shortest period of time, and migration to a new application is performed in a stepwise manner.