Service Level Data Storage Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large data storage systems are costly and complex, requiring a tradeoff between storage capacity and performance, with varying costs for additional storage capacity and performance upgrades, and existing methods for allocating storage resources are complex, error-prone, and costly, especially when application requirements change.
Innovation Solution
A data storage system that uses a plurality of physical storage devices with different performance metrics, where chunks of data storage capacity are associated with service levels indicative of performance, allowing for dynamic allocation and provisioning of storage resources based on service levels rather than specific components, enabling users to request storage in terms of capacity and performance without specifying particular hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If storage capacity is increased using slower devices, then cost is reduced, but performance deteriorates
Solution Approach 1:
The storage system is segmented into multiple service levels (e.g., performance-critical, standard, capacity-optimized) with different performance characteristics. Each service level is served by different storage devices or device types, allowing the system to offer both high performance and high capacity options without requiring all storage to be high-performance, thus reducing overall cost while maintaining performance where needed.
Solution Approach 2:
Different portions of the storage system are assigned different quality levels based on local requirements. High-performance storage devices are allocated to service levels requiring fast access, while lower-performance devices serve capacity-oriented service levels. This local differentiation allows the system to optimize cost-performance tradeoffs at each location rather than uniformly across the entire system.
2Speed
If storage system performance is increased, then cost increases
Solution Approach 1:
The system dynamically allocates storage resources to service levels based on actual performance requirements and usage patterns. Rather than provisioning all storage at the highest performance level, the system adapts resource allocation to match actual demands, allowing performance to be increased only where and when needed, thus controlling costs while maintaining necessary performance levels.
3Reliability
If storage resources are pre-provisioned to meet peak requirements, then performance requirements are satisfied, but cost increases due to unused capacity
Solution Approach 1:
Storage resources are designed to serve multiple service levels and applications simultaneously. The same physical storage infrastructure supports both performance-critical and capacity-optimized workloads by dynamically assigning resources to different service levels based on current needs. This multi-functionality allows the system to meet peak performance requirements without dedicating separate over-provisioned resources for each scenario, reducing overall cost.
4Reliability
If manual allocation of storage resources is performed, then specific performance requirements can be met, but complexity and error-proneness increase
Solution Approach 1:
The storage system automatically performs service level determination and resource allocation without requiring manual intervention. The system self-services by monitoring performance requirements, determining appropriate service levels, and allocating storage resources accordingly. This automation eliminates the complexity and error-proneness of manual allocation while maintaining reliable performance matching through systematic, rule-based decision-making.
Solution Approach 2:
The system incorporates feedback mechanisms that continuously monitor storage usage, performance metrics, and service level requirements. Based on this feedback, the system dynamically adjusts resource allocation to maintain performance requirements while optimizing cost. The feedback loop enables automatic adaptation to changing conditions without manual intervention, reducing complexity while ensuring reliable performance satisfaction.
Data Source
AI summary
The data storage capacity of a storage array, data center or networked data storage system is managed and allocated in terms of chunks of capacity at different service levels, where each service level is defined based on one or more of tiered storage policy settings, drive size, drive speed, drive count, RAID protection, engine fractions, bandwidth and availability and characterized by one or more performance capabilities, e.g., IOs per second. The physical storage devices at each service level may have similar capabilities or be tiered arrangements of devices having different capabilities. A request for storage indicates number of chunks and service level required, thereby avoiding typical allocation complexity. Monitoring and billing logic enables procurement on a per chunk basis at each service level, thereby optionally decoupling the cost of additional units of storage from underlying resources from the perspective of a customer.