Is Structured Cabling Dead?

In the past year or so, a lot of noise has been generated about wireless LANs. In fact, today if you were to set up a very basic wireless network, you could have it up and running in minutes, and that too at minimal cost. Perhaps that's why it's found to be more suitable than first laying down the cables. The speed at which wireless is gaining popularity is no joke, and ditto for the technology enhancements in this field. One hot topic that's emerging from all this is whether wireless networks can actually replace structured-cabling ones. While we'll see a lot of wireless deployments happening, we'll not see them replacing structured cabling in any way. Wireless will continue to extend existing wired networks and, in case of new networks, it will become a part of the overall architecture that will predominantly be cabled. Let's approach this argument logically by comparing the benefits and drawbacks of both. 

The trouble with wireless



Wireless networks are in a position where cabled networks were eight to ten years ago. 10 Mbps networks were popular at that time and were shared architectures based on hubs, meaning the 10 Mbps bandwidth was shared by all computers connected to the hub. Some deployments of 100 Mbps were also there, which were again hub-based. Let's look at wireless networks today. The 802.11b is the most popular standard with 802.11g catching up. While the first is able to deliver up to 11 Mbps of bandwidth, the latter takes it up to 54 Mbps. Also, some vendors such as Cisco, D-Link and Netgear have managed to double the throughputs to 108 Mbps if their equipment is used. While these are the theoretical maximum speed limits, they're never really achieved in practice. This is primarily due to RF signal losses and the build

quality of antennae in various wireless devices. Therefore, the throughput that's actually available is

much lower. 

Wired networks managed to evolve from shared to switched architecture, which dramatically improved their performance. In switched technology, everybody gets dedicated bandwidth, whether it's 10 or 100 Mbps. Unfortunately, Wireless LANs still use a shared architecture like the earlier hub-based wired networks, which is where the problem lies. If it's an 802.11b wireless access point, then 11 Mbps is shared by everyone connecting to it. There is something called a wireless LAN switch technology that's recently come up, which is becoming popular in enterprise wireless deployments. The outcome of this technology is a switch-based product that acts as the central connecting point for all wireless access points on the network. To explain how this works, consider the current scenario where all access points you deploy are connected to your existing layer 2 switches. This makes the manageability of access points a major problem. Plus, you have to work out a separate security structure for them by setting up a separate authentication server for them. Many wireless access points today are coming with integrated authentication services, but the problem here is managing the security on multiple access points. A WLAN switch is supposed to take away most of the features and functionality of these access points and centralize them. So, all wireless access points would terminate at the switch, which would then have all the security and manageability features built into it. 

While this sort of functionality is great for WLANs and intends to resolve a lot of problems associated with management and security, the main problem of shared architecture remains. Every access point still offers shared bandwidth. Moreover, all access points still connect to the switch using cables. If we look at the other side of the picture, then the shared architecture may not really be a limitation in some cases. If all you want to do is basic file sharing or access the Internet for Web browsing and e-mail, you don't need a lot of bandwidth. What's offered by wireless today is sufficient for that. But the needs of most organizations today go beyond that. A simple thing such as transferring a large file between a client and server over wireless can take a long time on a wireless network. Most organizations are running so many other bandwidth-intensive applications on a network, which a wireless network won't be able to take care of. While technically it's possible to design wireless networks to cater to such higher bandwidth requirements, the cost for doing that would be similar to doing the same over cable. In this case it's recommended to use cabling, which is not only a more mature technology but also offers much more bandwidth to start with and can easily be scaled up to give even higher bandwidth. Today, 10/100 Mbps switched bandwidth is common even at the desktop, Gigabit Ethernet has also picked up, and 10 Gbps has already been tested. So, there's a very well defined, high bandwidth path for cabled networks for the future. 

Where wireless shines



The hottest application for wireless networks is still mobility. If you have a lot of users with notebooks, then they would find wireless connectivity a boon. It's still good for extending a wired network to such mobile users. It's also useful for places where you have to quickly set up a network and then maybe dismantle it, such as in an exhibition. Another application that's already picked up across the world, and is slowly gaining ground in India, is wireless public hotspots. So you could go to a coffee shop with your notebook and surf the Net over a cup of coffee. The Barista chain of coffee shops has already started deploying WiFi. Cyber cafes can also benefit from WiFi deployments because the only application there is Internet access. WiFi at home is another application that's likely to pick up. 

The bottom line



In the medium term, going completely wireless is not a feasible option in place of cabling, with a few exceptions. Also, you can't completely ignore it, as it's becoming a part of every 'cabled' network infrastructure. 

Anil Chopra