Zero-downtime Migration

Cover from Zero-downtime Migration

Date: 14th July, 2008
By:
Format: White Paper

Download (subject to terms)
To access this paper you must first register and then login

We live in an increasingly 24 x 7 world. Both customers and internal users want access to relevant applications at any time of the day or night on any day of the year. In order to service this requirement companies need to implement policies that will ensure continuous availability rather than the merely high availability that has previously been the norm.

High availability ensures that applications will continue to run whenever there is a failure of some sort. This has historically allowed for systems to be taken down for maintenance. That is, high availability systems allow for unplanned outages but assume that planned outages are okay. However, for mission critical applications this is not okay: systems must continue to run through both planned and unplanned outages. This is what continuous availability provides.

In terms of day-to-day running, continuous availability can be provided through the use of cluster or grid technology that allows individual servers to be taken offline in a graceful manner, as and when required. Unfortunately, the same cannot be typically said of major migrations such as application consolidation, database migrations, implementations of master data management hubs, or simply upgrading from one software version to another.

In this paper we argue the case for a zero-downtime migration methodology that will enable such migrations within the context of continuous availability. This means exactly what it says: migration with, in theory, no downtime. Of course the theory may not be attainable: we know of one site, for example, where there was four minutes of downtime because the user wished to change application servers however, the migration itself did not require any downtime. We will discuss different approaches to zero-downtime later in this paper.

Zero-downtime contrasts with the traditional big bang approach to migration that has been most widely used on an historic basis, whereby there is a cut over to the new system, typically over a weekend. An excellent example of exactly what can go wrong with such an approach was provided by the opening of Terminal 5 at Heathrow, with its huge direct (

Post a review?

We welcome constructive criticism on all of our published content. Your name will be published against this review after it has been moderated. We reserve the right to contact you by email if needed.

Star Rating 1-5 (low-high):

If you don't want to see the security question, please register and login.