Cloud Native Adaptive Job Scheduler Framework
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud computing architectures face challenges in maintaining quality of service due to simultaneous dynamic loads from multiple customers, leading to service degradation and potential disruptions, especially in multi-tenanted SaaS environments where job priority and frequency decisions can result in unanticipated resource demands.
Innovation Solution
The implementation of an ensemble model combining (S)ARIMA, ARMA, and ARIMA models to predict resource consumption for job execution, allowing for optimized scheduling and resource allocation based on historical trends, thereby avoiding over-provisioning and ensuring efficient job execution within cloud computing environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If standard job scheduling configurations are used in multi-tenanted cloud environments, then job execution simplicity is maintained, but service quality degrades due to unanticipated resource demands from multiple customers
Solution Approach 1:
The system performs preliminary resource consumption prediction using (S)ARIMA models before job execution. Historical resource usage data is analyzed to forecast future resource demands, allowing the scheduler to proactively allocate resources and prevent service degradation before it occurs.
Solution Approach 2:
The job scheduling system transitions from static standard configurations to dynamic adaptive scheduling. The scheduler continuously adjusts resource allocation based on real-time resource predictions and actual consumption patterns, enabling the system to adapt to varying workload demands from multiple tenants while maintaining service quality.
2Reliability
If resource over-provisioning is implemented to handle peak demands, then service quality is maintained, but resource utilization efficiency decreases
Solution Approach 1:
The system implements a feedback mechanism where actual resource consumption during job execution is monitored and fed back to update the (S)ARIMA prediction models. This continuous learning process improves prediction accuracy over time, enabling precise resource allocation that matches actual demand without over-provisioning, thus maintaining service quality while optimizing resource utilization.
Solution Approach 2:
The system dynamically changes resource allocation parameters based on predicted workload characteristics. Instead of fixed over-provisioned resources, the scheduler adjusts CPU, memory, and storage allocations according to predicted job requirements, achieving efficient resource utilization while maintaining service quality through adaptive parameter optimization.
3Measurement precision
If multiple prediction models are combined to improve accuracy, then resource prediction precision increases, but system complexity increases
Solution Approach 1:
The system merges multiple prediction approaches by combining (S)ARIMA statistical models with actual runtime consumption data. This hybrid approach integrates historical pattern recognition with real-time observations, achieving high prediction accuracy while managing complexity through a unified prediction framework that leverages the strengths of different modeling techniques.
Data Source
AI summary
A request to execute a recurring job is received by a cloud computing environment. Thereafter, available computing resources available to execute the job are determined based on historical resource consumption trends. A resource prediction for the job is then generated using an ensemble model ((S)ARIMA model) that combines an autoregressive moving average (ARMA) model and an autoregressive moving average (ARIMA) prediction models. The resource prediction characterizes resources to be consumed for successfully executing the job. Execution of the job can then be scheduled by the cloud computing environment based on the resource prediction and the available computing resources. Related apparatus, systems, techniques and articles are also described.