Chances are, the web site you're using when you read this blog is an 8.5.3 server running on a Linux virtual machine on a VMWARE ESXi host at a big data center in Dallas, Texas. But not always. For most of the day today, if you'd read my blog or been to my website, you were using a different 8.5.3 server running on a Windows 2008 virtual machine on a different VMWARE ESXi host at a smaller data center in Seattle, Washington.

Did you notice? Was anything missing?

Probably not. That's because built in to the Lotus Domino server 20+ years ago was magic technology for replication. For the last several years that replication has been even faster when the servers are configured in a cluster. Not many people consider using cluster replication across a wide area network, but it works just fine.

Overnight, while I slept soundly in my hotel room outside Philadelphia, a drive failed on my Dallas based server. Within two minutes, my DNS provider noticed that my web server wasn't responding and switched over to responding with the alternate server's IP address. It sent me an email but other than that, nobody noticed. Throughout the day, in between client meetings, I worked to get the drive back into shape on the primary server and when it did come on-line, the servers re-synched and the DNS provider switched back over. I'll bet you didn't notice that either.

Could I do this with Sharepoint or a LAMP solution? Sure. Would it be as easy as turning on a feature? No. Definitely not.

The Domino server continues to be the best back-end solution for site hosting I've found -- at least for the kinds of sites I build. This is why I stick with it even when IBM periodically decides to abandon it (most of the years from 2000-2010) or does twisted things like building an entirely different development model without documentation. So far, IBM's competitors have failed to build as good a back end product. If they ever do, people will flock to it in droves.