On 22 Jun 2008 14:24:35 -0700, [EMAIL PROTECTED] (Gibney, Dave) wrote: > The issue that needs addressing is: Why, when technology has reached our > current level, is ANY customer visible downtime acceptable? Those "other > components" could, with today's capability, be properly redundant and > designed/coded/implemented such that the "SYSTEM" is never down at the public > interfaces. > It happens, that after dark on a weekend is MY best time to do my finances. > Yet, that is often when the "system is undergoing maintenance. > And the folks parked in daylight at the call center couldn't even understand > my irritation. > > I know that my little site at a University can't afford 9 nines, but damn > it, the credit card companies should be able to afford it, their fees are > significant to both buyer and seller.
How much is each additional nine worth? Of course the solution for ubiquitous up time, is probably to have a distributed system that is designed to work when data centers go down. If you have multiple data centers spread around large geographic areas, then our cost analysis in giving one computer more nines changes. This wasn't an option when the costs of each data center and the costs of data synchronization between data centers was prohibitive, but times change. We are looking at floods, earthquakes, terrorism making the concept of a single data center obsolete for large enterprises. So instead of spending money on 9 nines, spend it on more data centers. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
