Replication Resource Limits in Asymmetric Multi-Tenant Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In asymmetric multi-tenancy environments, existing data storage systems face challenges in managing replication resources efficiently, leading to resource shortages and performance degradation as the number of customers and data replication requests increase, as conventional mechanisms fail to enforce replication resource limits effectively.
Innovation Solution
Implementing capacity and stream limits at the target storage system, configured by a service provider administrator, to ensure that each tenant does not consume disproportionately high resources, with features like capacity soft and hard quotas, and stream limits that enforce resource usage through replication contexts and FSMO configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional replication mechanisms are used in multi-tenant environments, then data replication functionality is provided, but replication resources are not effectively managed leading to resource exhaustion and performance degradation
Solution Approach 1:
The patent applies parameter changes by introducing configurable replication resource limits (capacity limits and stream limits) that dynamically constrain replication operations. These limits are set as parameters in replication contexts and FSMO configurations, allowing the system to adjust replication behavior based on available resources and tenant requirements, thereby preventing resource exhaustion while maintaining data protection
Solution Approach 2:
The patent introduces an intermediary mechanism in the form of replication contexts and FSMO (File System Management Object) configurations that mediate between replication requests and actual data replication operations. These intermediaries enforce resource limits by intercepting and controlling replication streams, ensuring that no single tenant can consume excessive resources while still allowing legitimate replication operations to proceed
2Ease of operation
If replication resource limits are enforced through replication contexts and FSMO configurations, then fair resource utilization is achieved across tenants, but system complexity increases
Solution Approach 1:
The patent segments the replication control mechanism into distinct components: replication contexts that define resource limits for specific replication operations, and FSMO configurations that provide system-wide defaults and policies. This segmentation allows administrators to manage complexity by configuring limits at appropriate levels without having to control every individual replication operation manually
Solution Approach 2:
The patent creates universal configurations through FSMO defaults that can be applied across multiple replication contexts and tenants. By establishing system-wide default limits and policies in FSMO configurations, the same resource management rules can be automatically applied to numerous replication operations, reducing the administrative burden of configuring each tenant individually while maintaining fair resource utilization
Data Source
AI summary
A method for replicating data from a plurality of storage systems to a target storage system. The method implemented at the target storage system receives a replication request to replicate data from a source storage system to the target storage system, where the target storage system stores data replicated from the plurality of source storage systems. In response to the replication request, the target storage system identifies a replication resource limit associated with the data to be replicated from the source storage system. The target storage system then determines whether a replication resource requested by the replication request has exceeded the replication resource limit. The target storage system thus allows the replication resource to replicate data from the source storage system, in response to the replication resource not exceeding the replication resource limit.