CIOs and IT managers are chairman of selectors and coach in the data centre game, says Unisys New Zealand's Steve Griffin.
Experience is a brutal teacher, but it gets results.
When the All Blacks won the Rugby World Cup in 2011 they fielded a team with 709 caps. South Africa in 2007 had 688; England, in 2003, 638. Last year, against France, the All Black team boasted 853.
So while shiny new talent has its appeal, it’s experience that translates potential into performance. So it is with data centres and in choosing your provider, you become chairman of selectors.
The government’s drive for infrastructureas- a-service (IaaS) has resulted in a proliferation of new data centres with supply is outstripping demand.
Several organisations, some new, some seasoned, have made significant investment and expanded rapidly to meet what, at first blush, appears to be a single client’s potential requirements.
The market’s fascination with new and shiny things is not without risks, especially when trying to provide appropriate levels of resilience for mission critical systems.
A mature data centre (DC) is well beyond the initial problems that can beset complex facilities in their early days of operation.
This maturity brings well-proven architecture and tried and tested operational and engineering practices.
However, maturity should not be confused with age. If a DC gets consistent technology refreshes and upgrades to its mechanical and electrical infrastructure, its capability and reliability will be fully commensurate with newer facilities.
To provide disaster recovery capability and assure business continuity, a DC must operate geographically separate sites to mitigate the potential impact of natural disasters.
It must have not only high quality technology, but also a proven ability to deliver and support mission-critical business systems.
At the very least, a highly effective DC must be designed with an active-active configuration, with proven levels of historical uptime that exceeds the Uptime Institute’s guidelines for traditional Tier 2 and on a par with Tier 3 facilities.
To meet industry standards the redundant site infrastructure capacity components should have an expected availability of 99.741%.
The site should have multiple independent distribution paths serving the IT equipment, and all of its IT equipment must be dual-powered and fully compatible with the topology of the site’s architecture.
Further, it should have a concurrently maintainable site infrastructure with expected availability of at least 99.982%. A data centre’s internal processes and procedures should be in-line with the ITIL principles using the ISO20000 guidelines.
Additionally, they should be certified for quality processes under the ISO9001 guidelines and also be certified for Information Security under the ISO27001 certification.
A good data centre is not only carrier neutral and carrier diverse, it will also provide termination points for major carriers, elevating its status and improving the service it receives from them. You don’t achieve these standards with technology alone.
A DC provider does this by entrusting the management of its data centre to its own people to be responsible for your 'up time'.
An in-house maintenance team has an intimate knowledge of its own systems. They understand your DC’s unique setup. And they live locally and are on-call, 24x7. That’s a services oriented approach, one that runs counter to the offering of fully automated, ‘lights out’ facilities.
Management needs to carefully consider how to provide resilience for business critical systems.
The drive to meet the New Zealand Government’s move to IaaS, has raised the stakes. You, like Steve Hansen, have to make the big calls.