Centralized Fault Tolerance Handling for Software-Defined Data Center Services
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing maintenance mode handling techniques for services in software-defined data centers do not scale well and require redundant logic across multiple services, necessitating an understanding of the underlying infrastructure, which can lead to disruptions and inefficiencies.
Innovation Solution
A centralized management component evaluates fault tolerance information published by services to determine whether to place a host in maintenance mode, using data stores like PAAS or CAAS objects to abstract maintenance mode handling logic and avoid disruptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If individual services implement their own maintenance mode handling logic, then services can maintain control over their fault tolerance, but the system complexity increases and scalability deteriorates
Solution Approach 1:
The patent extracts maintenance mode handling logic from individual services and consolidates it into a centralized management component. This central component receives maintenance mode requests, evaluates fault tolerance information from all services, and makes centralized decisions about host maintenance mode placement, thereby reducing individual service complexity while maintaining system-wide reliability control
Solution Approach 2:
The centralized management component serves as a universal interface for all services regarding maintenance mode handling. It provides a single point of control that can evaluate fault tolerance for any service on any host, making the system more scalable without requiring each service to implement its own maintenance logic
2Reliability
If services implement redundant maintenance mode handling logic, then individual service reliability is maintained, but the overall system efficiency decreases due to redundant operations
Solution Approach 1:
The patent merges redundant maintenance mode handling logic from multiple services into a single centralized management component. This consolidation eliminates duplicate evaluation operations while maintaining the same reliability guarantees for each service, thereby improving overall system efficiency without sacrificing individual service reliability
3Reliability
If services require understanding of underlying infrastructure to implement maintenance mode logic, then services can make informed decisions, but the barrier to entry increases and operational complexity worsens
Solution Approach 1:
The centralized management component acts as an intermediary between services and the underlying infrastructure. Services publish their fault tolerance information to the central component without needing to understand infrastructure details. The central component uses this information to make maintenance mode decisions, thereby preserving informed decision-making while eliminating the need for services to directly understand complex infrastructure operations
Data Source
AI summary
The disclosure provides an approach for fault tolerance handling. Embodiments include determining, by a management component, that a host stores data relating to a service. Embodiments include receiving, by the management component, fault tolerance information from the service, the fault tolerance information comprising first information about host failures tolerated by the service and second information about existing host failures related to the service. Embodiments include determining, by the management component, based on the fault tolerance information from the service, whether the service will tolerate the host becoming unavailable. Embodiments include performing, by the management component, one or more actions based on the determining of whether the service will tolerate the host becoming unavailable.