Does anyone in this conversation know of a specific five-nines offering their business currently pays for? I don't really see it that much out in the wild.
Not pays for, except in the sense that it's a common demand/expectation made by "requesters" (read "management") to providers. Or it's a claim made by design teams as they propose and build solutions.
That was the original source of frustration that lead to me writing the article.
We have an effort of 3 and 4 nines when it comes to service. But any major outage blows that out of the water. Beyond that a large institution has so many points of failure that just a small office being offline could cut your numbers if your thresholds are offset. In the never ending battle of management and the number crunchers you are always faced with the expectations outweighing the capabilities.
Blame drive thru's for this instant and always on of an idea that gluttonous users can not get over.
5 9's of availability takes planning, effort, and cost commitment. But its not a myth. Tandem provided fault tolerant servers and a hot patchable OS. Now Stratus (and similar through NEC) are the main players in fault tolerant servers. http://www.stratus.com/Products/ftServerSystems
Oracle is making a play in the hot patchable OS, by applying ksplice. The technology is out there, though it still requires that entire systems, including personnel, are structured around it. In some cases, the cost is matched by value. In many cases its not.
Agreed. The only reason I label it a "myth" (besides it being an attention-grabbing title) is the almost slavish devotion to making EVERYTHING 5-9's reliable in an organization, from the servers to the internet connection to the coffee pot.
Well sure. I mean, certain systems are worth ANY expense!
I don't argue your basic position, since you're right and make a perfectly valid point, but I'd point out that less than 5 minutes of downtime is quite possible, given the right conditions.
I spent ten years working for Tandem Computers (now HP NonStop) and I can attest to more than one production system that had less than 5 minutes of downtime in a year. Consider those core, critical systems like 911, and ATM networks where downtime is not permitted...
There is where you find Tandem Systems where there is no single point of failure either in the hardware or the software, and barring human error (usually the cause of that lost 5 minutes,) mostly they don't go down.
SolarWinds solutions are rooted in our deep connection to our user base in the THWACK® online community. More than 150,000 members are here to solve problems, share technology and best practices, and directly contribute to our product development process. Learn more today by joining now.