Autonomic Characterization of Workloads Using Workload Fingerprinting

Khanna, Rahul; Ganguli, Mrittika; Narayan, Ananth; R, Abhiram; Gupta, Piyush

doi:10.1109/ccem.2014.7015482

Cited by 7 publications

(5 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our experiment we selected 9 events from the list of all performance events supported by the processor. The 9 events were selected from the results published by Khanna et al [5] Four Benchmarks were selected from the SPEC CPU 2006 suite [6] -a compute intensive application (POVRay), 2 memory intensive applications (mcf, Omnet++), and a streaming application (LBM). Each Benchmark was executed on an independent Virtual Machine (VM) simultaneously.…”

Section: Viexperimental Methodology and Resultsmentioning

confidence: 99%

“…As defined in our previous work [5], phase prediction uses the labeled data (with cluster information) used in the training example to build a phase-transition matrix that contains the likelihood of transitioning to another phase (or remain in the same phase). This data is used to predict the upcoming phase(s) and proactively determine the future demand based on that phase.…”

Section: B Phase Predictionmentioning

confidence: 99%

“…However in an IAAS (infrastructure as a service) environment like Amazon EC2 and Google GCEU multitenant environments the allocation of resources is based on the concept of shared resources and may only have a maximum SLO depending on CPU speed and not on workload phase based variable performance. In our previous work [5] we demonstrated the finger printing mechanism that allows us to forecast resource demands and tune system resources using phase-assisted dynamic characterization. In this paper we expanded our work by utilizing the collection of operational phases to match the nodes where they can be co-hosted.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Analytic technique for optimal workload scheduling in data-center using phase detection

Gupta

Koolagudi

Khanna

et al. 2015

5th International Conference on Energy Aware Computing Systems &Amp; Applications

Self Cite

View full text Add to dashboard Cite

Typically, complex resource-interdependence and heterogeneous workload patterns can result in sub-optimal job allocation leading to performance loss or under-utilization of compute resources. A well behaved model can anticipate the demand patterns and proactively react to the dynamic stresses in a timely and well optimized manner. For a workload hosting environment, pool of available resources are optimally configured and utilized to sustain certain expectation of Quality-of-Service (QoS) in the presence of power, thermal and reliability constraints. The workload (or job) scheduling mechanism is expected to withstand dynamic variations in demand stresses while maximizing the resource utilization and minimizing the performance loss. Furthermore, workloads can be co-allocated to the clusters with least amount of resource contention. In this paper we introduce the methodology that facilitates the coordinated scheduling of the workloads to the systems with least contentious resources through phase-assisted dynamic characterization. We describe the method to perform optimal job scheduling by using phase model synthesized by learning and classifying the run-time behavior of workloads.

show abstract

Section: Viexperimental Methodology and Resultsmentioning

confidence: 99%

Section: B Phase Predictionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Analytic technique for optimal workload scheduling in data-center using phase detection

Gupta

Koolagudi

Khanna

et al. 2015

5th International Conference on Energy Aware Computing Systems &Amp; Applications

Self Cite

View full text Add to dashboard Cite

show abstract

“…To track the phases automatically, Sherwood et al presented an architecture specifically for executed code (Sherwood, Sair, and Calder 2003). Also, some techniques based on machine learning have been proposed (Khanna et al 2014;Bhattacharyya, Sotiriadis, and Amza 2017;Jandaghi, Bhattacharyya, and Amza 2018). For example, to detect workload phases for inspiring more accurate resource provisioning, (Berral, Wang, and Youssef 2020a) proposed combining conditional restricted Boltzmann machines and distance-based clustering (i.e., k-means) to discover behavioral phases from resource usage metrics for auto-scaling.…”

Section: Related Work Phase Abstractionmentioning

confidence: 99%

Optimizing IT FinOps and Sustainability through Unsupervised Workload Characterization

Yang,

Arora,

Jha

et al. 2024

AAAI

View full text Add to dashboard Cite

The widespread adoption of public and hybrid clouds, along with elastic resources and various automation tools for dynamic deployment, has accelerated the rapid provisioning of compute resources as needed. Despite these advancements, numerous resources persist unnecessarily due to factors such as poor digital hygiene, risk aversion, or the absence of effective tools, resulting in substantial costs and energy consumption. Existing threshold-based techniques prove inadequate in effectively addressing this challenge. To address this issue, we propose an unsupervised machine learning framework to automatically identify resources that can be de-provisioned completely or summoned on a schedule. Application of this approach to enterprise data has yielded promising initial results, facilitating the segregation of productive workloads with recurring demands from non-productive ones.

show abstract

“…Khanna et al [29] describe a method for autonomic characterisation of workloads using workload fingerprinting. The authors focus on cloud computing and the need for the orchestration layer to forecast changing workload conditions in order to be able to meet Service Level Agreements (SLA).…”

Section: Previous Workmentioning

confidence: 99%

Autonomic Performance Optimization of Big Data Workloads

Genkin¹

View full text Add to dashboard Cite

show abstract

Autonomic Characterization of Workloads Using Workload Fingerprinting

Cited by 7 publications

References 22 publications

Analytic technique for optimal workload scheduling in data-center using phase detection

Analytic technique for optimal workload scheduling in data-center using phase detection

Optimizing IT FinOps and Sustainability through Unsupervised Workload Characterization

Autonomic Performance Optimization of Big Data Workloads

Contact Info

Product

Resources

About