Dynamic Resource Allocation for Cloud Tenants
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional static allocation of resources in cloud services leads to inefficient usage and higher costs due to over-allocation of resources, as tenants are often allocated the entirety of a core even if they do not utilize the full capacity, limiting the number of tenants that can be consolidated on a server.
Innovation Solution
Implementing dynamic allocation of rate-based computing resources, where tenants reserve a percentage of resources such as CPU, disk I/O, and network I/O, allowing for flexible allocation based on workload demands, with options for static or elastic reservations to accommodate steady or bursty workloads, and incorporating metering technologies to differentiate between insufficient demand and resource deprivation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If static allocation of resources is used, then resource reservations are guaranteed, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent implements dynamic resource allocation that transitions from static to dynamic assignment. The system continuously monitors resource usage and reassigns resources based on current demand, allowing the allocation to adapt over time. This resolves the contradiction by maintaining reliability through monitoring and enforcement mechanisms while improving productivity through flexible reassignment of underutilized resources to tenants who need them.
Solution Approach 2:
The system changes the parameter of resource allocation from fixed (static) to variable (dynamic). By introducing time-varying allocation parameters that adjust based on workload demands, the system can guarantee minimum resource reservations while allowing excess resources to be dynamically reallocated, thereby improving overall utilization efficiency without compromising reservation guarantees.
2Productivity
If more tenants are consolidated on a server, then cost efficiency improves, but resource allocation complexity increases
Solution Approach 1:
The patent implements self-service mechanisms where the resource allocation system automatically monitors, detects, and reassigns resources without manual intervention. The system autonomously manages the complexity of allocating resources among multiple tenants by implementing automated monitoring of resource usage patterns and dynamic reassignment protocols, thereby enabling high tenant consolidation while managing allocation complexity through automation rather than manual processes.
3Productivity
If dynamic allocation is implemented, then resource utilization efficiency improves, but difficulty of detecting and measuring resource usage increases
Solution Approach 1:
The patent implements comprehensive feedback mechanisms that continuously monitor resource usage by each tenant and provide real-time information to the allocation system. This feedback loop enables the system to detect actual resource consumption patterns, measure utilization accurately, and use this information to make informed dynamic allocation decisions. The feedback system resolves the measurement difficulty by establishing structured monitoring points and metrics that automatically track resource usage throughout the system.
Data Source
AI summary
Described herein are technologies relating to computing resource allocation among multiple tenants. Each tenant may have a respective absolute reservation for rate-based computing resources, which is independent of computing resource reservations of other tenants. The multiple tenants vie for the rate-based computing resources, and tasks are scheduled based upon which tenants submit the tasks and the resource reservations of such tenants.