Cost-aware computing

The Computing Continuum is the idea that increasingly fast networks enable substantial access to distributed compute resources, including the cloud, high performance computers (HPC), and the edge. Fluid computing across the continuum allows scientists and users to best utilize the resources that they have access to by pushing compute tasks to all resources rather than just one at a time. While novel methods and technologies have put this dream within reach, substantial hurdles are still necessary to cross in order to put this in the hands of of compute users everywhere.

The Scalable Cost-Aware Cloud Infrastructure Management and Provisioning (SCRIMP) project aimed to address these challenges by developing new, more efficient cloud provisioning methods and integrating these new methods into automated cloud and HPC access tools. Now, the Distributed Execution of Lambdas with Tradeoff Analysis (DELTA) project encompasses the cloud-specific methodologies developed through SCRIMP along with cutting edge techniques using serverless computing to push compute everywhere in the computing continuum.

Our research focuses on four core areas: (1) developing profiling methods to characterize workload execution characteristics; (2) exploring tradeoffs in computational cost across computing resources, from cloud to edge, in order to inform intelligent resource selection; and (3) creating an automated cost-aware provisioning service using serverless infrastructure to deploy tasks on diverse computing resources based on time, monetary, and service cost constraints.

Further reading

  • N. Hudson, H. Khamfroush, M. Baughman, D.E. Lucani, K. Chard, I. Foster, “QoS-aware edge AI placement and scheduling with multiple implementations in FaaS-based edge computing”. Future Generation Computer Systems, 2024.

  • R. Ananthakrishnan, Y. Babuji, M. Baughman, J. Bryan, K. Chard, R. Chard, B. Clifford, I. Foster, D. Katz, K. Hunter Kesling, C. Janidlo, R. Mello, L. Wang, “Enabling Remote Management of FaaS Endpoints with Globus Compute Multi-User Endpoints”. PEARC, 2024.

  • M. Baughman, N. Hudson, I. Foster, K. Chard, “Balancing Federated Learning Trade-Offs for Heterogeneous Environments”. PerCom, 2023.

  • S. Caton, M. Baughman, C. Haas, R. Chard, I. Foster, K. Chard, “Assessing the Current State of AWS Spot Market Forecastability”. SuperCompCloud, 2022.

  • M. Baughman, I. Foster, K. Chard “Exploring tradeoffs in federated learning on serverless computing architectures”. eScience, 2022.

  • O. Almurshed, P. Patros, V. Huang, M. Mayo, M. Ooi, R. Chard, K. Chard, O. Rana, H. Nagra, M. Baughman, I. Foster, “Adaptive edge-cloud environments for rural AI”. International Conference on Services Computing, 2022.

  • M. Baughman, I. Foster, K. Chard, “Enhancing Automated FaaS with Cost-aware Provisioning of Cloud Resources”. eScience, 2021.

  • M. Baughman, R. Kumar, I. Foster, K. Chard, “Expanding Cost-Aware Function Execution with Multidimensional Notions of Cost”. 1st Workshop on High Performance Serverless Computing. 2021.

  • R. Kumar, M. Baughman, R. Chard, Z. Li, Y, Babuji, I. Foster, and K. Chard, “Coding the Computing Continuum: Fluid Function Execution in Heterogeneous Computing Environments”. 30th Heterogeneity in Computing Workshop. 2021.

  • M. Baughman, N. Chakubaji, H. Truong, K. Kreics, K. Chard, and I. Foster, [“Measuring, Quantifying, and Predicting the Cost-Accuracy Tradeoff”] (https://acris.aalto.fi/ws/portalfiles/portal/41076847/paper.pdf) 3rd IEEE International Workshop on Benchmarking, Performance Tuning, and Optimization for Big Data Applications. 2019.

  • C. Wu, T. Summer, Z. Li, A. Woodard, R. Chard, M. Baughman, Y. Babuji, K. Chard, J. Pitt, and I. Foster, [“ParaOpt: Automated Application Parameterization and Optimization for the Cloud”] (https://labs.globus.org/pubs/wu-cloudcom-2019.pdf) IEEE International Conference on Cloud Computing Technology and Science. 2019.

  • M. Baughman, S. Caton, C. Haas, R. Chard, R. Wolski, I. Foster, and K. Chard, [“Deconstructing the 2017 changes to AWS spot market pricing”] (https://labs.globus.org/pubs/Baughman_deconstructing_2019.pdf) 10th Workshop on Scientific Cloud Computing. 2019.

  • M. Baughman, R. Chard, L. Ward, J. Pitt, K. Chard, and I. Foster, [“Profiling and Predicting Application Performance on the Cloud”] (https://labs.globus.org/pubs/Baughman_UCC_2018.pdf) 11th IEEE/ACM International Conference on Cloud Computing. 2018.

  • M. Baughman, C. Haas, R. Wolski, I. Foster, and K. Chard, “Predicting Amazon spot prices with LSTM networks” 9th Workshop on Scientific Cloud Computing. 2018.

  • R. Chard, K. Chard, R. Wolski, R. Madduri, B. Ng, K. Bubendorfer, and I. Foster, “Cost-Aware Cloud Profiling, Prediction, and Provisioning as a Service,” IEEE Cloud Computing 4 (4) : 48-59. 2017.
  • R. Chard, K. Chard, B. Ng, K. Bubendorfer, A. Rodriguez, R. Madduri, and I. Foster, “An Automated Tool Profiling Service for the Cloud,” 2016 16th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (CCGrid), Cartagena, 2016, pp. 223-232.
  • R. Chard, K. Chard, K. Bubendorfer, L. Lacinski, R. Madduri and I. Foster, “Cost-Aware Cloud Provisioning,” e-Science (e-Science), 2015 IEEE 11th International Conference on, Munich, 2015, pp. 136-144.
  • R. Wolski, J. Brevik, R. Chard, and K. Chard. Probabilistic guarantees of execution duration for Amazon spot instances. The International Conference for High Performance Computing, Networking, Storage and Analysis. 2017