Two-Level Throttling for Multitenant Resource Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multitenant computing environments, tenants lack knowledge of resource usage by other tenants, leading to unfair allocation of throttling values, which can result in resource overconsumption by some hosts, impacting others.
Innovation Solution
Implementing a two-level throttling system with per-tenant and per-host throttling values, managed by a system administrator, where activity levels are monitored, and throttling is applied based on predefined service levels to ensure fair resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If per-tenant throttling values are assigned without per-host granularity, then resource allocation is simplified, but individual hosts within a tenant can overconsume resources and impact other tenants
Solution Approach 1:
The patent segments the throttling mechanism into two distinct levels: per-tenant throttling values and per-host throttling values. This segmentation allows the system to manage resources at both the aggregate tenant level and the individual host level, preventing any single host from overconsuming resources while maintaining simplified overall management through the dual-level structure.
2Reliability
If throttling values are set statically for each host, then resource allocation is predictable, but the system cannot adapt to changing activity levels and service requirements
Solution Approach 1:
The patent implements dynamic throttling by continuously monitoring activity levels for each host and comparing them against service level thresholds. When activity levels exceed predefined thresholds, the system automatically adjusts throttling values in real-time, allowing the system to adapt to changing workloads while maintaining predictable resource allocation through service level agreements.
Solution Approach 2:
The system incorporates feedback mechanisms by monitoring activity levels and using this information to dynamically adjust throttling values. The controller receives feedback about current resource consumption patterns and modifies throttling parameters accordingly, creating a closed-loop control system that maintains reliability while adapting to changing conditions.
3Object-generated harmful factors
If monitoring and dynamic throttling are implemented for all hosts, then resource fairness is improved, but system complexity and computational overhead increase
Solution Approach 1:
The patent reduces complexity by segmenting the monitoring and control functions into distinct per-tenant and per-host components. This segmentation allows the system to apply monitoring and dynamic throttling selectively at appropriate levels, improving resource fairness while managing complexity through modular architecture rather than requiring centralized control of all hosts simultaneously.
Data Source
AI summary
Computer-implemented methods for throttling in a multitenant environment are provided. Aspects include assigning, by a system administrator of the multitenant environment, a service level for each tenant of a plurality of tenants of the multitenant environment and obtaining a throttling value for each of a plurality of hosts associated with each of the plurality of tenants in the multitenant environment. Aspects also include monitoring, by a controller of the multitenant environment, an activity level for each of the plurality of tenants of the multitenant environment. Based on a determination that the activity level for one of the plurality of tenants has reached a maximum value specified by the service level, aspects include throttling one or more of the plurality of hosts associated with the one of the plurality of tenants based on the throttling value for the one or more of the plurality of hosts.