Staggered Server Farm Reboot Cycles for Continuous Access
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for performing maintenance operations in server farms, such as rebooting and workload transfer, often result in connection or service outages, disrupting user access to hosted resources and potentially causing data loss or boot storms that overload infrastructure.
Innovation Solution
A method that configures a server farm preference with a primary and secondary farm, allowing for seamless session transfer and staggered reboot cycles to minimize disruptions, ensuring continuous user access and reducing the risk of outages during maintenance operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If maintenance operations are performed on server farms, then system reliability is improved, but service continuity deteriorates
Solution Approach 1:
The patent segments the server farm into multiple independent server farms, each capable of hosting the same hosted resources. This segmentation allows maintenance operations to be performed on one server farm while others continue to serve users, thereby maintaining service continuity while improving system reliability through distributed architecture.
Solution Approach 2:
The patent implements preliminary action by pre-configuring server farm preferences for hosted resources to include multiple server farms as options. Before maintenance occurs, the system is prepared with alternative server farms that can take over hosting responsibilities, enabling seamless failover and minimizing service disruption during maintenance operations.
2Reliability
If reboot operations are performed on machines, then system reliability is improved, but boot storms are caused
Solution Approach 1:
The patent segments the reboot operation into individual machine-level operations rather than simultaneous farm-wide reboots. Each machine can be rebooted independently at different times, preventing boot storms while still achieving the reliability benefits of periodic reboots through distributed, staggered execution.
Solution Approach 2:
The patent implements periodic action by scheduling reboot operations at different times for different machines rather than performing them simultaneously. This periodic, staggered approach allows system reliability to be improved through regular maintenance while avoiding the harmful effects of boot storms that occur when all machines reboot at once.
3Productivity
If workload transfer is performed, then resource optimization is improved, but connection outages occur
Solution Approach 1:
The patent segments the workload transfer process into gradual, incremental transfers rather than abrupt, simultaneous migrations. By distributing the transfer over time and across multiple server farms, the system achieves resource optimization while minimizing connection outages and maintaining continuous user access during the transition.
Solution Approach 2:
The patent uses server farm preferences as an intermediary mechanism that mediates between the source and destination server farms during workload transfer. This preference system allows hosted resources to be progressively reassigned to different server farms, enabling resource optimization while preventing connection outages through controlled, preference-based routing during the transition period.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
One or more aspects of this disclosure may relate to using a configurable server farm preference for an application, desktop or other hosted resource. Additional aspects may relate to moving server farm workloads based on the configurable server farm preference. Further aspects may relate to performing reboot cycles, a reboot schedule and on-demand rebooting. Yet further aspects may relate to staggering individual machine reboot operations over a specified period of time and performing reboot operations such that some machines are available for user sessions during a reboot cycle.