If you are responsible for managing servers from which you are physically separated, you should consider using a special card, usually in a PCI
Requires Free Membership to View
When you register you’ll also receive the latest news, advice and technical tips designed specifically for midmarket IT leaders like yourself. Our award-winning editorial team will give you immediate access to emerging business and technology trends.
Scot Petersen, Editorial Director, SearchCIO-MidmarketEach of the server manufacturers has a different name for their card. For example, IBM refers to its "Remote Supervisor Adapter" card and Dell calls theirs a "DRAC," but the basic functions are the same. The cards have their own Ethernet connection which can operate as long as there is power, and you use them by establishing a terminal session to this card. Many also have a keyboard/video/mouse connection too, so you can use that from a remote KVM switch.
Aside from making remote administration easier, these cards have several added security benefits. For example, they separate user traffic from admin traffic, which can be a big security benefit if your network is set up for "out-of-band" management. That is, you can isolate your management traffic to a separate VLAN or switch so that users can't get there, without affecting any connectivity to the server's primary network connection. Also, this is a security benefit because your operating system doesn't have to also be listening on any management ports, like ssh or Telnet, or for Windows Terminal Server or PCAnywhere, so obviously, your system won't be vulnerable to exploits in those services.
If your servers are a name brand like IBM, Dell or HP, you're probably better off getting a remote console card from the server manufacturer, so that it integrates with the BIOS. If you've rolled your own server from various components, you should probably go with a board like the eRIC, which acts like a stand-alone computer.
This was first published in October 2005