Pre-deployed Agents for Zero-Downtime Enterprise Manager Upgrades
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Upgrading enterprise management systems is a time-consuming and resource-intensive process, often resulting in significant downtime and monitoring loss due to the need to replace multiple collection agents across various network nodes, and maintaining backward compatibility can be costly and constrain the development process.
Innovation Solution
A pre-deployment technique where a second set of agents is activated to collect data from targets and send it to a new management server, while the first set of agents continues to send data to the old server, allowing for incremental and out-of-place activation of new agents, enabling zero or near-zero downtime upgrades even if the new system is not backward compatible.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the enterprise manager is completely shut down before deploying the new system, then the upgrade can be performed, but significant downtime and monitoring loss occur
Solution Approach 1:
The patent applies preliminary action by pre-deploying the new enterprise manager system alongside the existing system before the old system is shut down. This allows the new system to be ready and tested in advance, enabling a seamless transition that minimizes downtime and avoids monitoring loss during the upgrade process.
2Loss of time
If backward compatibility is maintained to reduce downtime, then monitoring loss is reduced, but development costs increase and the development process is constrained
Solution Approach 1:
The patent applies segmentation by separating the old and new enterprise manager systems into distinct, parallel implementations. This allows the new system to be developed independently without maintaining backward compatibility constraints, while the old system continues to operate independently, enabling cost-effective development of the next generation system.
Solution Approach 2:
The patent uses an intermediary approach by introducing a dual-system architecture where both old and new systems coexist temporarily. This intermediary state allows data to be collected by both systems simultaneously, enabling validation of the new system without requiring backward compatibility, thus reducing development costs while maintaining operational continuity.
3Adaptability or versatility
If all collection agents are replaced with new agents, then the new system features are activated, but the complexity of managing the upgrade process increases significantly
Solution Approach 1:
The patent applies self-service by enabling the new enterprise manager system to automatically collect and validate data from collection agents independently of the old system. This self-validating approach simplifies upgrade management because the new system can autonomously verify its functionality and data integrity without requiring complex manual coordination of all agent replacements.
Solution Approach 2:
The patent maintains continuity of useful action by having both old and new collection agents operate simultaneously during the transition period. This allows the new system to be activated incrementally while the old system continues to function, reducing the complexity of managing complete agent replacement and enabling gradual validation of the new features.
Data Source
AI summary
Techniques are described for upgrading systems that include a first set of agents for collecting target data from a plurality of targets and sending the target data to a first management server. In an embodiment, a second set of agents is pre-deployed. Each pre-deployed agent is configured to, upon activation, collect target data from a respective target of the plurality of targets and send the target data to a second management server. While the pre-deployed agent remains inactive, an agent from the first set of agents collects and sends the target data to the first management server. When a first batch of one or more pre-deployed agents is activated, each activated agent begins collecting target data and sending the target data to the second management server. For agents not in the first batch, the first set of agents continues to collect and send the target data.