Redefining mission impossible
What does ‘mission critical’ mean in 2011? For years, companies selling high availability hardware would base their specifications on the percentage of uptime that they could guarantee. The holy grail was ‘five nines’; 99.999%, which equates to just under minutes of downtime per year. However, systems that could guarantee such uptime were incredibly expensive.
In an era of cloud computing and easy connectivity, do such measurements apply any more? Years ago, it was difficult to get one server to failover to another via a simple dynamic DNS management service. Cloud computing concepts like infrastructure-as-a-service and platform-as-a-service lacked maturity. Today, these are becoming more accessible.
In this new world, it becomes easier for IT departments to swap computing loads between their own systems, and those of third party providers. In the meantime, more reliable hardware has fallen in price, and it is possible to pick up full tolerant servers at relatively low costs for deployment not only in head offices, but in branch offices too.
None of this makes downtime go away, but it commoditizes uptime, and switches the focus elsewhere. The onus is now on IT departments to explain what they are doing for their internal IT customers in a more business-friendly way. Discussing the nuances of mission criticality and high availability systems does not serve the line of business customer. Instead, they want to know whether they will be able to continue working at times of peak demand, and rely on their systems not to go down.
Instead of making promises of complete uptime that are impossible to keep, perhaps it would be more productive for IT departments to rely on these cheap, reliable solutions, and then schedule their maintenance downtime during periods when business customers do not need systems to be available. A call center that only works from 9 AM to 6 PM may need to handle spikes in computing demand during events when the volume of calls might escalate. But, it may be perfectly acceptable for the system to be down for hours at a time at night.
Instead of quantifying system uptime, then, by dwelling on the number of nines, it makes more sense in 2011 to focus on what a business really needs–and then allocating sufficient budget to get there, rather than boiling the ocean with arcane and expensive technical solutions.
Credit: © Nmedia - Fotolia.com