Storage Area Network Agent Load Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional storage area network management applications face inefficiencies due to agents inundating the management device with concurrent requests, leading to timeouts and poor system performance, primarily caused by insufficient bandwidth and processing power.
Innovation Solution
Implementing a mechanism where agents calculate a current load value to determine whether to forward requests to the remote management entity, thereby preventing overburdening and optimizing request forwarding based on the entity's capacity, using weight values and threshold load values to manage the burden.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multiple agents concurrently send requests to the management device, then the system can gather configuration information from multiple switches in parallel, but the management device becomes inundated with requests leading to timeouts and poor performance
Solution Approach 1:
The agent performs preliminary actions by calculating the current load value before forwarding requests to the management device. This advance assessment allows the agent to determine whether the management device can handle additional requests, preventing request inundation while maintaining efficient parallel operation across multiple agents
Solution Approach 2:
The system implements feedback mechanisms where agents monitor the management device's response to previous requests and adjust their request forwarding behavior accordingly. By tracking whether the management device responds within timeout windows, agents can dynamically control request rates, ensuring reliable operation while maximizing productivity
2Speed
If the management device processes all requests simultaneously, then the system responds quickly to agent inquiries, but the management device lacks sufficient processing power and bandwidth to handle concurrent requests
Solution Approach 1:
The agent applies partial action by selectively forwarding only those requests that the management device can currently handle within its processing capacity. By calculating the current load value and comparing it against thresholds, the system performs just enough parallel processing to maintain speed while avoiding excessive requests that would overwhelm the management device's limited processing power
3Reliability
If agents resend requests after timeouts, then the system ensures information is retrieved, but the management device experiences further congestion and wasted processing time
Solution Approach 1:
The agent applies preliminary anti-action by calculating the current load value and assessing whether the management device is likely to respond within the timeout window before actually sending the request. This preventive measure stops requests that would likely timeout, eliminating the need for resending and avoiding further congestion on the management device while maintaining reliable information retrieval
Data Source
AI summary
A management control center communicates with multiple agents in a storage area network. On behalf of the management control center, the agents communicate with a switch resource management entity that maintains configuration information about resources such as a group or groups of switches in the storage area network. Based on a chain of communications from the management control center through the agents to the remote management entity, the management control center can indirectly learn about a configuration of switch resources in the storage area network. The agent utilizes a calculated current load value (e.g., a value indicating a relative burden placed on the remote management entity to service the forwarded set of requests) to selectively forwards requests to the switch resource management entity. This technique ensures that the switch resource management entity is not overwhelmed by a task of servicing too many requests at the same time.