Zero Downtime Migration

Upgrading systems can still make good business sense, but often only if zero or very little system downtime can be assured during the migration.

The banking and financial crisis has triggered an unprecedented global downturn that has heralded an era of turbulence and change. Businesses are intensely focused on two commercial priorities; retaining customers and maintaining financial controls. Excellent customer service, brand differentiation, and operational efficiency are no longer aspirational ‘nice to haves’ but hardcore business imperatives. And achieving these goals is dependent on one thing — the zero-downtime of business-critical systems and applications.

In today’s ‘always-on’ multi-channel business climate, system downtime has become unacceptable. Globalisation and the Internet has created a 24x7 world where products and services are increasingly only available online, customer service expectations are high, and loyalty is fickle. The proliferation of online shopping, customer call centres, online travel reservations and banking mean business reputations now stand or fall against service levels in an instant.

The continuous operation of critical systems, maintaining the continuous availability of applications and their underlying databases, is paramount to maintaining operational excellence and market leadership.

Counting the cost

Unplanned outage represents significant tangible cost to a business. Research undertaken by the Warwick Business School in 2008 highlights the growing dependency of many UK business sectors on IT infrastructures. The research revealed retail is particularly at risk, with a probable downtime cost of over £350,000 per hour – with considerably higher stakes for global online players like Google or Amazon — compared to finance which faces downtime costs of over £100,000 per hour.

For online retail brands like Amazon the impact of one hour’s downtime can mean the loss of planned and impulse purchases, loss of traffic from affiliates, and the possibility of customers buying from the next closest competitor, all spurred on by service non-availability.

Who dares wins

Organisations still need to migrate to new technologies or hardware platforms, and to upgrade existing environments to improve performance and overall functionality. Investment to maintain competitive leadership and guarantee exemplary customer service makes good business sense, but often only if zero or very little system downtime can be assured during a major migration project.

The ‘big bang’ method of rolling out new systems – often implemented over the weekend in the hope that everything will be resolved by Monday – exposes companies to considerable risk. The stakes are even higher for organisations that need systems to be available and operational 24x7, such as a billing application in the telecommunications sector. Here, planned downtime is simply not an option.

The challenge is to remove the risk in the migration process for business critical systems and — even if migration is not on the horizon — mitigate against the impact of an unplanned outage for continuous business operations.

Achieving zero-downtime

Major migrations are complex tasks that typically take months, and often years. They are also prone to failures and overruns, especially with respect to the data elements of those migrations. According to research conducted by Bloor Research, only 13% of data migration projects are brought in on time and on budget. There are many reasons for this: lack of experienced practitioners, inadequate project scoping and overconfidence to name but a few. Many companies panic at even the thought of a migration project, particularly where large volumes of data are concerned, and projects are often shelved or postponed for this very reason. However, postponements tend only to compound the problem and may end up necessitating a migration, not from version x to x+1, but to version x+n, which can result in even greater complexity! If there really is a valid business reason for a migration to version x+1, technical considerations should not be allowed to stand in the way. For the implementation and execution of migration projects, companies therefore need a clear and reproducible methodology - one that will make such projects appear less daunting.

Page 2 - overcoming bi-directional replication hurdles