As businesses increasingly rely on computers by the thousands, flung far and wide around the globe, the job of managing themthat is, monitoring, identifying, and correcting problemshas become an industry of its own. Enterprise management is a hot expression in the computer industry now. Very pricey software systems have been developed to centrally monitor computers, networks, hubs, routing hardware, UPSs, and even computer room fire alarms. These systems detect problems and can notify staff via pager, email, printouts, or, who knows, probably even carrier pigeon. The purpose of these systems is to catch problems as they develop, with any luck, before they disrupt people trying to do their work. Instrumentation is the key here: Equipment has to be designed to be monitored. A TCP/IP-based protocol called Simple Network Management Protocol (SNMP) has been around for years, and "managed" network equipment is capable of being probed and reconfigured via SNMP. Along with this capability comes a hefty price tag, but the net cost of maintaining and dispatching staff to fix problems is much greater. My small LAN with four users and a handful of development and online servers doesn't need a $20,000 management system, managed hubs, and the like. But, even in my little office, I find myself constantly checking to make sure the servers are up, that they have plenty of disk space, and that the Internet connection is working. What I really want is something to check these things periodically and let me know whether something's amiss. I guess plenty of other people do, too. Free enterprise is a wonderful thing. I searched the Web and found a handful of packages targeted for small LANs just like mine. If you have a LAN you depend on for your business, you might want to check them out. Hearing about a problem from your pager is a lot nicer than hearing about it from a client or an employee! Using these products, you can specify a series of computers or devices to be periodically tested. The tests can include ping, SNMP, file sharing services, Windows Service activity, disk space availability, server responsiveness, and so on. Failures can be announced to a list of alert recipients via pager, email, or printout. (Different problems may call for different announcement methods, of course. If your LAN is down, an email alert won't get delivered.) Some products can even send announcements to selected employees based on their work schedules. Using these tools, you can enter a list of your most important network servers and other resources, and rest assured that if something goes wrong with any of them, you'll be notified immediately. The following are a few products worth investigating. Each of these programs can detect whether a remote server is active, can test various types of network services (for example, file sharing, Web services, and so on), and can send a message to a pager if a failure occurs.
Reviewing these products, I found that the product quality is roughly proportional to the price, and it's best, if possible, to use one that runs as a Windows service so that it can do its job even while nobody is logged in to the monitoring computer. These products are not quite as sophisticated or well designed as their $20,000 cousins, but they might be just the ticket for a small business. |