We rent two servers at 1and1.com to host our clients' websites. One of them just passed a major milestone: being up without reboot for over one year. Just a few minutes ago, it's uptime statistics were: 372 days, 12 hours, 39 minutes, the other reports being up 214 days, 22 hours, 17 minutes. So why aren't all our websites up 100% of the time? Well, this leads to a discussion of what parts of the digital pipeline between you and your website are reliable what percent of the time. We employ a monitoring service that checks the availablity of several of our sites from different places around the world. One of the sites we host, for instance, has a reported downtime of 5 minutes in this past month. But the server's been running along just fine for the past nine months? What went down?
Well, it was the network going to the server had a glitch, in all likelyhood. The ugly truth in the modern internet is that the least reliable link in the whole process is the network itself. The server stays up and running, the end users are there all the time, but the network sometimes has problems. Remember that the Internet grew out of the DARPANET project from the U.S. Department of Defense. The design goal was a network that could withstand a nuclear attack. So why does this amazing network sometimes have connectivity problems? Lots of reasons: lightening strikes, power outages and, most infamous of all, the rogue backhoe digging where it shouldn't. What can I as your system administrator or you as a customer do about it? Unfortunately, not much. We live in a fallen world and we're lible to death and corruption, as are our network cables. Still, the uptime we're seeing on our sites, though not perfect, are still plenty good enough for us to have healthy, productive sites and businesses. Though not perfect, the uptime's plenty good enough. May the same be said of us all.