TAR Based Hotspot Prediction in Cloud Data Centres

TAR Based Hotspot Prediction in Cloud Data Centres

Anu Valiyaparambil Raveendran (Mahatma Gandhi University, Kottayam, India) and Elizabeth Sherly Sherly (Iiitmk, Kerala, India)
Copyright: © 2019 |Pages: 22
DOI: 10.4018/IJGHPC.2019070101
OnDemand PDF Download:
No Current Special Offers


In this article, the authors studied hotspots in cloud data centers, which are caused due to a lack of resources to satisfy the peak immediate requests from clients. The nature of resource utilization in cloud data centers are totally dynamic in context and may lead to hotspots. Hotspots are unfavorable situations which cause SLA violations in some scenarios. Here they use trend aware regression (TAR) methods as a load prediction model and perform linear regression analysis to detect the formation of hotspots in physical servers of cloud data centers. This prediction model provides an alarm period for the cloud administrators either to provide enough resources to avoid hotspot situations or perform interference aware virtual machine migration to balance the load on servers. Here they analyzed the physical server resource utilization model in terms of CPU utilization, memory utilization and network bandwidth utilization. In the TAR model, the authors consider the degree of variation between the current points in the prediction window to forecast the future points. The TAR model provides accurate results in its predictions.
Article Preview

1. Introduction

Now a day’s most online services hosted in data centre servers demand uninterrupted 24 × 7 services. This increases the scope of live migration and equally its complexity. In live migration, VM transfer take place when it is in working mode ideally with no down time. Live migration can be done in pre-copy, post copy or hybrid form of both methods as per demand. Total migration time, down time, amount of data transferred etc. are some key factors which determines the efficiency of VM migration process. Even though migration is an inevitable process in cloud computing, recurrent migration may degrade the performance of system and may lead to SLA violation.

A hotspot is an overloaded condition of physical machine (PM) where it does not have adequate resources to serve immediate demanded request. During the occurrence of hotspot the downtime of physical servers increases drastically which degrade the performance level of PM and may lead to SLA violation. If the resource utilization shows a consistent increase for some time window, then there is high probability to occur a hotspot in near future. Frequent creation of hotspots in physical servers affect QoS of entire system. It is because additional cooling systems has to be provided to cool the overloaded machines. So, handling hotspots are very essential especially when the data centre is handling high available or online services.

The following are the contribution of this work:

  • Interference aware live migration (IALM) is introduced to overcome interference issue in physical machine during secure live migration.

  • Interquartile method (IQR) is used to calculate the minimum threshold value to identify hotspot.

  • Trend aware regression (TAR) model is applied as a load prediction model for hotspot detection in cloud environment.

We organize the paper as follows. Section II narrates the motivation and Section III describes related work. In section IV explain the system model. In section V we present our prediction model and migration algorithm. In section VI we show the analysis of our prediction model. Finally, section VII concludes the article.


3. Literature Survey

Most of the research works in live migration optimization considers nature of current workload as the deciding factor for the occurrence of VM migration. In normal scenario, VM migration is initiated as soon as the host is found to be overloaded. VM migration is an expensive affair and sometime affect the destination machine in terms of interference, frequent VM migrations may lead to performance degradation and SLA violation. Both current and future workloads play competent role on VM migration performance. The ill effect of hotspot can be mitigated by predicting the future work load using a forecasting model. Hotspot is an unfavourable condition in physical machine where it cannot serve the immediate peak if request and lead to SLA violation. So, resource prediction model plays a significant role in hotspot avoidance and ensure relevant number of VM migration between the servers in cloud data centres.

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 14: 6 Issues (2022): Forthcoming, Available for Pre-Order
Volume 13: 4 Issues (2021)
Volume 12: 4 Issues (2020)
Volume 11: 4 Issues (2019)
Volume 10: 4 Issues (2018)
Volume 9: 4 Issues (2017)
Volume 8: 4 Issues (2016)
Volume 7: 4 Issues (2015)
Volume 6: 4 Issues (2014)
Volume 5: 4 Issues (2013)
Volume 4: 4 Issues (2012)
Volume 3: 4 Issues (2011)
Volume 2: 4 Issues (2010)
Volume 1: 4 Issues (2009)
View Complete Journal Contents Listing