Lucky network administrators get to run cabling in brand new buildings. When linking buildings, lucky network administrators with big budgets get to run fiber or maybe even invest in a microwave or laser connection. But not all network admins are lucky. For example, what do you do when you have to run new cabling in your 150-year-old building or link multiple buildings on a shoestring budget? In this article, I will show you ways in which you can use wireless access points to bridge both floors and buildings. Since when does wireless work building-to-building?
The guidelines state that 802.11b and 802.11a communications are limited in effective range. 802.11b connections are rated for a maximum of 300 feet, and most 802.11a connections aren't considered to be effective beyond 60 feet. Of course, the farther you get from the access point, the slower the connection. Therefore, you may think that the usefulness of wireless communications is limited. These limitations would imply that 802.11x connections can't handle building-to-building or floor-to-floor connections. However, as described on the O'Reilly Network, 802.11b communications have been possible at distances of up to 10 miles using an antenna featuring an empty Pringles potato chip can. Longer distances (up to 20 miles) have been achieved using either commercially available antennas or something a little more substantial than a Pringles can. With this type of range, it's easy to use 802.11b to link networks within buildings and even in different buildings without having to run fiber optic or potentially expensive T1 lines. Get to the point
You have two architectural choices to consider when using 802.11b to bridge between networks: point-to-point bridging or point-to-multipoint bridging. Point-to-point bridging means exactly what it sounds like: Communications flow from one access point to another when connecting locations. Point-to-multipoint bridging is a little more involved. With point-to-multipoint, one central access point at your main location serves as a connecting point for all other locations/floors. The difference between point-to-point and point-to-multipoint is analogous to the difference between ThinNet Ethernet and 10Base-T. Like ThinNet, if an access point in a point-to-point configuration fails, communication across the network will be broken, while still allowing the computers connected on either side of the break to talk to each other. In a point-to-multipoint configuration, if the failing point is one of the multiple access points, only that point's computers will lose communication. However, if the central access point fails, all communication breaks down on the network. Because the cost of 802.11b gear is reasonable, and point-to-multipoint connections create the potential for a complete network-communications failure, it doesn't make sense to try to bring all the wireless building connections in on the same access point. If you have no choice but to deploy a point-to-multipoint connection, take care to consider bandwidth implications. Your total incoming connections bandwidth can't exceed the maximum bandwidth of the access point. According to the 802.11b standard, the maximum top speed on a wireless network is 11 megabits per second. Assuming that you have the remote connections locked down to a 1-Mbps connection speed, this means that, at most, your bandwidth would allow only 11 simultaneous connections to one access point. At this rate, each connection gets an equal slice of the bandwidth pie available from the central access point. If you want to use more than 11 connections, you would have to enable Quality of Service (QoS) on your central access point to ensure that each connection would achieve a specific connection speed. Because not all access points have the QoS feature, you must check to see if the feature exists on the access point you're considering before purchasing it. If you don't want to lock down the speed of connections between access points, you can ensure the central access point doesn't become flooded by allowing it to negotiate the connection speeds on its own. The problem with this method is that the closer connections will obtain the faster speeds, which could cause congestion when other connection points have a demand for a higher bandwidth.