Online Spike Detection in Cloud Workloads

Amardeep Mehta; Jonas Dürango; Johan Tordsson; Erik Elmroth

doi:10.1109/IC2E.2015.50

Online Spike Detection in Cloud Workloads

Amardeep Mehta, Jonas Dürango, Johan Tordsson, Erik Elmroth

Research output: Chapter in Book/Report/Conference proceeding › Paper in conference proceeding › peer-review

253 Downloads (Pure)

Abstract

We investigate methods for detection of rapid workload increases (load spikes) for cloud workloads. Such rapid and unexpected workload spikes are a main cause for poor performance or even crashing applications as the allocated cloud resources become insufficient. To detect the spikes early is fundamental to perform corrective management actions, like allocating additional resources, before the spikes become large enough to cause problems. For this, we propose a number of methods for early spike detection, based on established techniques from adaptive signal processing. A comparative evaluation shows, for example, to what extent the different methods manage to detect the spikes, how early the detection is made, and how frequently they falsely report spikes.

Original language	English
Title of host publication	[Host publication title missing]
Publisher	IEEE - Institute of Electrical and Electronics Engineers Inc.
Pages	446-451
Number of pages	6
DOIs	https://doi.org/10.1109/IC2E.2015.50
Publication status	Published - 2015
Event	2nd IEEE Workshop on Cloud Analytics - Tempe, AZ, United States Duration: 2015 Mar 12 → …

Conference

Conference	2nd IEEE Workshop on Cloud Analytics
Country/Territory	United States
City	Tempe, AZ
Period	2015/03/12 → …

Subject classification (UKÄ)

Control Engineering

Free keywords

cloud
cloud workload
workload spike
spike detection

Access to Document

10.1109/IC2E.2015.50

iwca15

1 Licentiate Thesis

Control Strategies for Improving Cloud Service Robustness
Dürango, J., 2016 Jun 14, Department of Automatic Control, Lund Institute of Technology, Lund University. 111 p.
Research output: Thesis › Licentiate Thesis

Open Access
File

1 Finished

EIT_VR CLOUD Cloud Control
Papadopoulos, A. V. (Researcher), Robertsson, A. (Researcher), Åkesson, J. (Researcher), Årzén, K.-E. (Researcher), Maggio, M. (Researcher), Dellkrantz, M. (Researcher), Tärneberg, W. (Researcher), Li, Z. (Researcher), Dürango, J. (Researcher) & Kihl, M. (Researcher)
2013/01/01 → 2016/12/31
Project: Research

Cite this

@inproceedings{89c1b8a4719740d997a70fb09987bb3f,

title = "Online Spike Detection in Cloud Workloads",

abstract = "We investigate methods for detection of rapid workload increases (load spikes) for cloud workloads. Such rapid and unexpected workload spikes are a main cause for poor performance or even crashing applications as the allocated cloud resources become insufficient. To detect the spikes early is fundamental to perform corrective management actions, like allocating additional resources, before the spikes become large enough to cause problems. For this, we propose a number of methods for early spike detection, based on established techniques from adaptive signal processing. A comparative evaluation shows, for example, to what extent the different methods manage to detect the spikes, how early the detection is made, and how frequently they falsely report spikes.",

keywords = "cloud, cloud workload, workload spike, spike detection",

author = "Amardeep Mehta and Jonas D{\"u}rango and Johan Tordsson and Erik Elmroth",

year = "2015",

doi = "10.1109/IC2E.2015.50",

language = "English",

pages = "446--451",

booktitle = "[Host publication title missing]",

publisher = "IEEE - Institute of Electrical and Electronics Engineers Inc.",

address = "United States",

note = "2nd IEEE Workshop on Cloud Analytics ; Conference date: 12-03-2015",

}

TY - GEN

T1 - Online Spike Detection in Cloud Workloads

AU - Mehta, Amardeep

AU - Dürango, Jonas

AU - Tordsson, Johan

AU - Elmroth, Erik

PY - 2015

Y1 - 2015

N2 - We investigate methods for detection of rapid workload increases (load spikes) for cloud workloads. Such rapid and unexpected workload spikes are a main cause for poor performance or even crashing applications as the allocated cloud resources become insufficient. To detect the spikes early is fundamental to perform corrective management actions, like allocating additional resources, before the spikes become large enough to cause problems. For this, we propose a number of methods for early spike detection, based on established techniques from adaptive signal processing. A comparative evaluation shows, for example, to what extent the different methods manage to detect the spikes, how early the detection is made, and how frequently they falsely report spikes.

AB - We investigate methods for detection of rapid workload increases (load spikes) for cloud workloads. Such rapid and unexpected workload spikes are a main cause for poor performance or even crashing applications as the allocated cloud resources become insufficient. To detect the spikes early is fundamental to perform corrective management actions, like allocating additional resources, before the spikes become large enough to cause problems. For this, we propose a number of methods for early spike detection, based on established techniques from adaptive signal processing. A comparative evaluation shows, for example, to what extent the different methods manage to detect the spikes, how early the detection is made, and how frequently they falsely report spikes.

KW - cloud

KW - cloud workload

KW - workload spike

KW - spike detection

U2 - 10.1109/IC2E.2015.50

DO - 10.1109/IC2E.2015.50

M3 - Paper in conference proceeding

SP - 446

EP - 451

BT - [Host publication title missing]

PB - IEEE - Institute of Electrical and Electronics Engineers Inc.

T2 - 2nd IEEE Workshop on Cloud Analytics

Y2 - 12 March 2015

ER -

Online Spike Detection in Cloud Workloads

Abstract

Conference

Subject classification (UKÄ)

Free keywords

Access to Document

Fingerprint

Research output

Control Strategies for Improving Cloud Service Robustness

Projects

EIT_VR CLOUD Cloud Control

Cite this