InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Estimating Bandwidth for Data Replication of Application based on Filesystem and undelying Storage Disk

IP.com Disclosure Number: IPCOM000219518D
Publication Date: 2012-Jul-05
Document File: 5 page(s) / 132K

Publishing Venue

The IP.com Prior Art Database


Disclosed is a method for Estimating the Bandwidth for Data Replication of an Application, based on File System and rate of change of data on the underlying storage disk

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 48% of the total text.

Page 01 of 5

Estimating Bandwidth for Data Replication of Application based on Filesystem and undelying Storage Disk

1.1 Abstract

Evaluating the bandwidth for replication of Application data through Storage Replication technologies, is a very manual and cumbersome process as it exists today. Most methods collect data at a 'host end' to estimate rate of change of data at the disk level. This is then put through a painstaking manual process of identifying the required file system's volumes, identifying the data writes on the disks of the volumes and then estimating the bandwidth for the required data replication.

This paper proposes a method to automate the process identifying data writes for a given file system from a storage end, which makes it easier to estimate the

bandwidth for a given set of application data in a repeatable manner as well as maintain a history of such activity for the purpose of trending and capacity planning

1.2 Introduction

The most important requirement for estimating bandwidth required for data replication of an application is identifying the rate of change of data. In countries

where cost of bandwidth is high, it is important that the right estimation is done on the bandwidth thereby reducing the risk of under estimating or over estimating the bandwidth, which could result in an impact to either performance or cost.

Existing methods to do this are mostly manual and require a loads of effort. Most methods use the 'iostat' or similar data from the server side, to identify the

writes that happen at the disk level. Since the 'iostat' (or similar tools) provides data at a server level, this dataneeds to be mapped to the storage volumes and disks to identify the writes a particular file system or mount point is making on the storage. Once this data is identified, this is then used for estimation of the bandwidth.

However, in this method, there is no easy way to maintain a history or to trend and project future growth unless more manual effort is expended.

1.3 Background and Related Work

Solutions from different vendors that exist today provide guidance on estimating bandwidth once the rate of change of data has been identified.

The following links provide guidance on how bandwidth can be estimated, for different solutions as provided by individual vendors.



Page 02 of 5

http://read.virtualizeplanet.com/?p=91 http://searchstoragechannel.techtarget.com/feature/Network-bandwidth-Sizing-for-long-distance-storage-networking-applications

In essence, the method offered is as follows:

1. Identify the data to use for bandwidth estimation - this is the difficult and time consuming part

2. Use a formula to estimate bandwidth using identified data - repeatable but manual; accuracy of measurement depends on the accuracy of data gathered

which can vary with every iteration

This paper proposes a way to automate the process of identifying the rate of change of data on the storage...