Efficient Autoscaling in the Cloud Using Predictive Models for Workload Forecasting

Nilabja Roy, Abhishek Dubey, Aniruddha S. Gokhale

IEEE International Conference on Cloud Computing, CLOUD 2011, Washington, DC, USA, 4-9 July, 2011 2011

PDF DOI

Why This Matters

Cloud computing promises cost-effective computing through dynamic resource allocation, yet static provisioning based on average load results in either performance degradation during peaks or resource waste during valleys. This work innovates by integrating workload prediction with control-theoretic optimization to make dynamic scaling decisions that satisfy performance requirements while minimizing operational costs. The predictive approach anticipates load changes rather than reacting to them, enabling faster and more efficient resource adjustments.

What We Did

This paper presents efficient autoscaling algorithms for cloud environments that predict future workload and allocate resources to maintain quality of service while minimizing operational costs. The work develops a model predictive control approach that uses performance models to estimate required resources based on predicted workload. It combines multiple techniques including workload forecasting using autoregressive moving average models, response time analysis, and resource allocation optimization to determine optimal machine allocation.

Key Results

Experiments with realistic workload traces demonstrate that the predictive algorithm determines effective resource allocations that allocate extra machines only when predicted load increases. The approach successfully satisfies quality of service objectives while achieving significant cost savings compared to static or reactive approaches. The algorithm shows effective performance across different cost function configurations and workload patterns.

Full Abstract

Cite This Paper

@inproceedings{Roy2011a,
  author = {Roy, Nilabja and Dubey, Abhishek and Gokhale, Aniruddha S.},
  booktitle = {IEEE} International Conference on Cloud Computing, {CLOUD} 2011, Washington, DC, USA, 4-9 July, 2011},
  title = {Efficient Autoscaling in the Cloud Using Predictive Models for Workload Forecasting},
  year = {2011},
  acceptance = {22.4},
  pages = {500--507},
  abstract = {Large-scale component-based enterprise applications that leverage Cloud resources expect Quality of Service(QoS) guarantees in accordance with service level agreements between the customer and service providers. In the context of Cloud computing, auto scaling mechanisms hold the promise of assuring QoS properties to the applications while simultaneously making efficient use of resources and keeping operational costs low for the service providers. Despite the perceived advantages of auto scaling, realizing the full potential of auto scaling is hard due to multiple challenges stemming from the need to precisely estimate resource usage in the face of significant variability in client workload patterns. This paper makes three contributions to overcome the general lack of effective techniques for workload forecasting and optimal resource allocation. First, it discusses the challenges involved in auto scaling in the cloud. Second, it develops a model-predictive algorithm for workload forecasting that is used for resource auto scaling. Finally, empirical results are provided that demonstrate that resources can be allocated and deal located by our algorithm in a way that satisfies both the application QoS while keeping operational costs low.},
  bibsource = {dblp computer science bibliography, https://dblp.org},
  biburl = {https://dblp.org/rec/bib/conf/IEEEcloud/RoyDG11},
  category = {selectiveconference},
  contribution = {colab},
  doi = {10.1109/CLOUD.2011.42},
  file = {:Roy2011a-Efficient_Autoscaling_in_the_Cloud_Using_Predictive_Models_for_Workload_Forecasting.pdf:PDF},
  keywords = {cloud computing, autoscaling, workload prediction, resource allocation, performance modeling, quality of service},
  project = {cps-middleware},
  tag = {platform},
  timestamp = {Wed, 16 Oct 2019 14:14:54 +0200},
  url = {https://doi.org/10.1109/CLOUD.2011.42}
}
Quick Info
Year 2011
DOI 10.1109/CLOUD.2011.42
Keywords
cloud computing autoscaling workload prediction resource allocation performance modeling quality of service
Research Areas
scalable AI ML for CPS
Search Tags

Efficient, Autoscaling, Cloud, Predictive, Models, Workload, Forecasting, cloud computing, autoscaling, workload prediction, resource allocation, performance modeling, quality of service, scalable AI, ML for CPS, 2011, Roy, Dubey, Gokhale