Sherlock: Lightweight Detection of Performance Interference in Containerized Cloud Services

Joshi, Kartik; Raj, Arun; D, Janakiram

doi:10.1109/hpcc-smartcity-dss.2017.68

Cited by 6 publications

(3 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Heifer leverages a lightweight performance model to predict the performance of MapReduce applications based on the online measured resource utilization and the captured VM interference. Although the method of modeling performance interference can help to accurately find the best node for an instance, Kartik Joshi [36] states that the performance of containers running in co-located VMs would be unpredictable when some VMs excessively consume the shared hardware resources. It thus proposes a lightweight subscriber-centric mechanism named Sherlock to detect performance interference and it defines a metric named IScore to estimate its impact on cloud services.…”

Section: Related Workmentioning

confidence: 99%

Be Good Neighbors: A Novel Application Isolation Metric Used to Optimize the Initial Container Placement in CaaS

et al. 2020

View full text Add to dashboard Cite

Many cloud providers benefit from fine-grained resource management in Container as a Service (CaaS), e.g., energy savings. However, while optimizing power consumption, cloud providers also need to consider service performance. Few studies in the literature performed a balanced optimization of the two objectives in the initial container placement (ICP). As such, the obtained solutions are often locally optimal. In this work, we model the ICP as a bi-objective optimization problem. Unfortunately, directly applying a conventional performance modeling method in solving the problem is complex and costly, especially for a large-scale application scenario. It also fails to capture the impacts of some non-resource constraints such as high availability constraints. To tackle these issues, we propose a novel application isolation metric to quantify the overall service performance. With regard to the power consumption objective, we opt for a non-linear model since a linear one makes power consumption indistinguishable in a homogeneous setting. With the two building blocks, we introduce an optimization algorithm named First Fit based improving Genetic Algorithm (denoted by FF-based-IGA) to find the best ICP solution. Our experimental results are as follows: 1) the proposed application isolation metric is effective at improving service performance, and 2) FF-based-IGA achieves a desirable balance between power consumption and service performance.

show abstract

Section: Related Workmentioning

confidence: 99%

Be Good Neighbors: A Novel Application Isolation Metric Used to Optimize the Initial Container Placement in CaaS

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Joshi, Raj, and Janakiram developed a system called Sherlock aimed at the detection of performance interference in containerized cloud services [33]. The described system also works by profiling the user workload and detecting discrepancies from its internal model and performance measurements made on the running application.…”

Section: J Comparison With Other Monitoring Systemsmentioning

confidence: 99%

A Novel Approach to Client-Side Monitoring of Shared Infrastructures

2020

View full text Add to dashboard Cite

Modern computing is migrating towards more prevalent use of shared infrastructures and is transforming physical and fixed resources into a more dynamic and virtualized environment. This trend has proved to be a boon for advanced monitoring and profiling tools and approaches available for infrastructure operators and service providers. On the other end of the spectrum, the clients have been provided with previously unavailable flexibility and simple access to vast resources at the price of being unable to detect or correctly locate issues caused by lower layers of the new complex but abstracted infrastructure. This issue is most visible with performance degradation caused by other tenants of the same infrastructure, tenants formally invisible for other end-users. Although performance interference in shared infrastructures is a known problem that is covered by research aimed at detection and mitigation by infrastructure operators, there is a lack of research aimed at end-users and their ability to detect and locate such performance degradation. This paper covers the areas of shared networking, virtualized compute and grid compute infrastructures, gives an analysis of the specifics of each of the infrastructures, presents a novel approach for client-side monitoring and discusses the practical implementation issues. The system presented in the paper aims to provide end-users with valuable insight into underlying physical infrastructure without burdening them with complex and heterogeneous nature of monitored devices by reducing the shared infrastructure to a single representative generated node simply accessible as a generic SNMP (Simple Network Management Protocol) enabled networked device. The primary aim of the system is to enable end-users to detect and locate the domain of the issue causing the performance interference and not to replace or reimplement existing general-purpose monitoring systems. The system was tested on a regional e-Infrastructure within VI-SEEM and NI4OS-Europe projects where it performed as expected.

show abstract

“…Some authors propose mechanisms that help detect the presence of performance interference. Joshi et al [15] measure an application in controlled conditions, constructing a throughput-vs-utilization curve. In real deployment, significant departure from that curve is interpreted as a sign of performance interference.…”

Section: Related Workmentioning

confidence: 99%

Duet Benchmarking: Improving Measurement Accuracy in the Cloud

Bulej,

Horký,

Tůma

et al. 2020

Preprint

View full text Add to dashboard Cite

We investigate the duet measurement procedure, which helps improve the accuracy of performance comparison experiments conducted on shared machines by executing the measured artifacts in parallel and evaluating their relative performance together, rather than individually. Specifically, we analyze the behavior of the procedure in multiple cloud environments and use experimental evidence to answer multiple research questions concerning the assumption underlying the procedure. We demonstrate improvements in accuracy ranging from 2.3× to 12.5× (5.03× on average) for the tested ScalaBench (and DaCapo) workloads, and from 23.8× to 82.4× (37.4× on average) for the SPEC CPU 2017 workloads.

show abstract

Sherlock: Lightweight Detection of Performance Interference in Containerized Cloud Services

Cited by 6 publications

References 19 publications

Be Good Neighbors: A Novel Application Isolation Metric Used to Optimize the Initial Container Placement in CaaS

Be Good Neighbors: A Novel Application Isolation Metric Used to Optimize the Initial Container Placement in CaaS

A Novel Approach to Client-Side Monitoring of Shared Infrastructures

Duet Benchmarking: Improving Measurement Accuracy in the Cloud

Contact Info

Product

Resources

About