Browse Prior Art Database

Method for Intelligent dynamic compression and adaptive change in sampling frequency of monitoring messages in clouds

IP.com Disclosure Number: IPCOM000245185D
Publication Date: 2016-Feb-18
Document File: 5 page(s) / 91K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to analyze the current cloud environment and dynamically introduce compression of monitoring messages and disable it whenever it is not required. A related method is also revealed by which monitoring data sampling frequency can be intelligently altered according to the situation

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.

Page 01 of 5

Method for Intelligent dynamic compression and adaptive change in sampling frequency of monitoring messages in clouds

Cloud market is becoming more and more demanding - so as the cloud services. One of the most important aspect of cloud is to have a stable monitoring system. In case of monitoring services - data is sent from each nodes to the central cloud controller or a separate monitoring server where in this data is further analyzed.

In this, controller henceforth means entity responsible for handling monitoring data received from monitored resources.

As new types of deployment and compute strategies are getting introduced to cloud , the instances that each compute host can support is also getting increased to more than what was before. As the number of instances increases - the quantity of data that is to be sent from compute hosts to controller also increases.

In case of container clouds - the number of instances can vary (sudden increase or decrease) dynamically as containers are short lived compared to traditional VMs and container density per host is much more (more than 10x) vis-a-vis VMs.

Resulting huge monitoring data may choke primary cloud facilities by consuming much needed bandwidth for other cloud operations and facilities. However compression of data may effect the performance especially when its not required (when there is enough network bandwidth)

Monitoring is usually done as sampling process, where at periodic intervals sample of a resource usage is collected, usually pre configured.

There are no existing solution to:

Dynamically enable compression of huge monitoring data when required and possible

Dynamically disable compression of huge monitoring data when not required or not possible

Dynamically alter sampling frequency of huge monitoring data and optimize to collect "just needed" data which can be handled without compromising any other cloud functions due to monitoring overheads.

based on current cloud environment

1


Page 02 of 5

This art aims to analyze the current cloud environment and dynamically introduce compression of monitoring messages and disable it whenever is not required. A related method is also revealed by which monitoring data sampling frequency can be intelligently altered according to the situation

Server Component which does intelligent compression and adaptation in sampling frequency

The central part of this article is Server Component which does intelligent compression based on current cloud details

This component relies on following fact to make the decision whether monitoring data is to be compressed or not


1) In order to enable compression - controller and compute should have enough CPU ( "enough CPU"-> could be made configurable).


2) Compression only required when the network bandwidth is less than needed ("needed" -> could be made configurable)


3) Intelligent adaptation of monit...