| Disaster Recovery Musings |
| Written by Eric Novikoff | |||||||||||||||||||||||||||||||
|
Here at ENKI we get a lot of questions about DR, probably since the AppLogic technology we base our services on already provides increased resistance to hardware failure, so customers think to themselves, "well, what would it take to tack another nine onto my reliability percentage?" The traditional wisdom in the IT industry is that each nine of reliability costs you double the amount of money. For example, going from 99.9% uptime to 99.99% uptime might double your cost because this transition typically requires fully redundant hardware. AppLogic bypasses these economics to some extent by giving you 3+ nines out of the box (if you know the tricks of the trade to get the most from it) while only requiring n+1 redundancy rather than 2*n. This also presupposes that you have fully redundant network hardware, though networking hardware typically reaches three 9's without redundancy. To discuss disaster recovery with customers, I have created a table that summarizes my views with respect to AppLogic versus non AppLogic-based options to increase reliability over the reference standard of private server hosting. It isn't absolute, but it's a great starting point for a discussion. One note: the DR site solutions listed assume that you aren't using AppLogic at the DR site. If these sites were deployed with AppLogic in a utility billed model such as ENKI uses, the costs could be reduced dramatically, since the resources allocated to them could be increased once the DR site was actually in use. Essentially you'd add the "AppLogic-based Hosting" column costs to your current site costs, with additional savings on labor because both sites could be administered through the same process.
Comments (0)
 Write comment
|
|||||||||||||||||||||||||||||||
Subscribe to this comment's feed
Show/Hide comments
Show/Hide comment form
| < Prev | Next > |
|---|