Application Power Signature Analysis

Hsu, Chung-Hsing; Combs, Jacob; Nazor, Jolie; Santiago, Fabián; Thysell, Rachelle; Rivoire, Suzanne; Poole, Stephen W

doi:10.1109/ipdpsw.2014.90

Cited by 10 publications

(3 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our literature search resulted in only a small amount of closely-related work. In [14], the authors apply clustering techniques on HPC power data to fingerprint applications. The authors rely on six time domain features and demonstrate that such features are enough to differentiate between the ten programs used for testing.…”

Section: High Performance Computingmentioning

confidence: 99%

Catch Me If You Can: Using Power Analysis to Identify HPC Activity

Copos,

Peisert

2020

Preprint

View full text Add to dashboard Cite

Monitoring users on large computing platforms such as high performance computing (HPC) and cloud computing systems is non-trivial. Utilities such as process viewers provide limited insight into what users are running, due to granularity limitation, and other sources of data, such as system call tracing, can impose significant operational overhead. However, despite technical and procedural measures, instances of users abusing valuable HPC resources for personal gains have been documented in the past [43], and systems that are open to large numbers of loosely-verified users from around the world are at risk of abuse. In this paper, we show how electrical power consumption data from an HPC platform can be used to identify what programs are executed. The intuition is that during execution, programs exhibit various patterns of CPU and memory activity. These patterns are reflected in the power consumption of the system and can be used to identify programs running. We test our approach on an HPC rack at Lawrence Berkeley National Laboratory using a variety of scientific benchmarks. Among other interesting observations, our results show that by monitoring the power consumption of an HPC rack, it is possible to identify if particular programs are running with precision up to 97% and recall of 95% even in noisy scenarios.

show abstract

Section: High Performance Computingmentioning

confidence: 99%

Catch Me If You Can: Using Power Analysis to Identify HPC Activity

Copos,

Peisert

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…More recently, Hsu et al developed power signatures based on short statistical summaries of applications' power traces [10]. However, they evaluated them only qualitatively, by visual inspection of the clusterings they yielded for a small dataset, and they did not attempt to compare them to results based on time-series metrics or signatures with more complex features.…”

Section: Related Workmentioning

confidence: 99%

“…We choose these two features as our initial signature, following the work of Hsu et al [10], which showed qualitatively that combining two location parameters seemed to yield good clusterings for power traces.…”

Section: B Feature-vector Representationmentioning

confidence: 99%

Power Signatures of High-Performance Computing Workloads

Combs

Nazor

Thysell

et al. 2014

2014 Energy Efficient Supercomputing Workshop

Self Cite

View full text Add to dashboard Cite

Workload-aware power management and scheduling techniques have the potential to save energy while minimizing negative impact on performance. The effectiveness of these techniques depends on the stability of a workload's power consumption pattern across different input data, resource allocations (e.g. number of cores), and hardware platforms.In this paper, we show that the power consumption behavior of HPC workloads can be accurately captured by concise signatures built from their power traces. We validate this approach using 255 traces collected from 13 high-performance computing workloads on 4 different hardware platforms. First, we use both feature-based and time-series-based distance metrics to cluster our traces, and we quantitatively show that feature-based clusterings segregate traces by workload just as effectively as the more compute-and space-intensive time-series-based clusterings. Second, we demonstrate that unlabeled traces can be classified by workload with over 85% accuracy, based only on these concise statistical signatures.

show abstract