Detecting Faulty Resources in Cloud Delivery Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Resource delivery systems face challenges in efficiently detecting and repairing faulty shared computing resources, as they often appear available but fail to respond to client connections, leading to poor user experiences and inefficient load balancing, with current methods being impractical due to the large number of resources and potential disruption to other users.
Innovation Solution
A system that uses targeted application probing, where a resource director identifies potentially faulty resources by configuring a client device to attempt connections with selected shared computing resources, determining faults based on access attempts, and taking corrective actions such as rebooting, using tags to filter and direct connection requests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional methods are used to detect faulty shared computing resources, then comprehensive detection can be achieved, but system disruption and user impact increase significantly
Solution Approach 1:
The patent segments the detection process by introducing a dedicated probe agent that operates independently from the main system. This probe agent specifically targets faulty resources for detection without requiring comprehensive system-wide probing, thereby isolating the detection activity from normal user operations and minimizing disruption.
Solution Approach 2:
The patent introduces a probe agent as an intermediary component that mediates between the system administrator and the shared computing resources. This probe agent performs detection activities on behalf of the administrator, enabling fault detection without requiring direct administrator intervention that would disrupt user access to resources.
2Measurement precision
If comprehensive resource probing is performed to identify faulty resources, then detection completeness improves, but detection time and system performance degrade
Solution Approach 1:
The patent applies partial action by having the probe agent focus only on probing resources that are suspected to be faulty, rather than performing comprehensive probing of all resources. This targeted approach achieves sufficient detection completeness for problematic resources while significantly reducing the time and computational overhead compared to exhaustive system-wide probing.
3Reliability
If faulty resources are not detected timely, then system stability is maintained, but user experience and load balancing efficiency deteriorate
Solution Approach 1:
The patent implements feedback mechanisms where the probe agent continuously monitors resource health and reports findings to the system administrator. This feedback loop enables timely detection of faulty resources, allowing the load balancer to adjust resource allocation dynamically and maintain high load balancing efficiency without compromising system stability.
Data Source
AI summary
One example disclosed method involves a computing system determining that a first shared computing resource, included among a plurality of shared computing resources managed by a controller, is potentially faulty. The system may configure the controller to identify the first shared computing resource with a tag, the tag configured to cause the controller, in response to a client requesting the controller for access to a first application, to assign the first shared computing resource to deliver the first application to the client. The system may instruct the client to request the controller for access to the first application. The system may determine that the first shared computing resource is faulty based on the client being unable to access the first application at a time of application of the tag to the first shared computing resource. The system may take a corrective action with respect to the first shared computing resource.