Services : Systems

System Administration Tasks

We do nightly backups to disk, for easy restores, and weekly backups to tape. We keep at least two month's worth of backup tapes available for restoration if necessary, more if requested.

Our systems are monitored in two ways: Exodus monitors our live server, and a service called Uptime, running at MIT in Cambridge, MA monitors all our systems. We will be notified via our Surepage pager if either service detects a problem.

Our web server (AOLserver) and our database (Oracle) are monitored by a suite of monitoring tools from ArsDigita. These tools are able to recognize some error conditions and automatically correct them, and will notify us via pager if an error occurs that they are not able to correct.

We run a link-verification suite an all sites we host on a monthly basis, to check for broken links. This is most important to sites which contain links to external sites, which may have moved or been taken down since the link was last checked.

Support Software

We provide two stats packages: Analog and NetTracker 4.0. Analog is a freeware product which does basic site statistics; it's very fast and accurate and provides most of the information most web site owners are interested in, but one thing it does not provide is a visitor count. NetTracker is on the opposite end of the spectrum - it's not overly fast, but it's very sophisticated, highly configurable, and provides all the information you could ever want (and then some). All our customers have Analog stats available to them; Nettracker is an extra-cost option.

We use AddWeb to do search engine submissions; we generally recommend that sites be submitted once a quarter or when the site has changed substantially. Most search engines specify a minimum time between submissions on the same site, and AddWeb observes this automatically.

The Future

We're currently building a high-availability, high-performance server cluster to provide a virtually 100% uptime web service (as close as you can get, anyway). Not that we have a lot of downtime now; so far we've had 100% uptime with our current systems, and it's hard to improve on that. But a server cluster provides a lot more redundancy; an entire system can fail and the web sites will stay up. That's a Good Thing!