In terms of Quality of Service (QoS), Ethernet/IP networks historically could be likened to a McDonald’s restaurant — they were relatively cheap and efficient, but network managers couldn’t expect a lot of bells and whistles. If managers wanted assured QoS for mission-critical applications, they could go upscale and run some ATM through their campus or WAN environment. Sure, it might have been a bit more expensive and slightly tougher to learn, but at least they knew their networks weren’t going to be
dropping packets all over the place.
Despite ATM’s superior QoS qualities, the technology never really took hold beyond the WAN and a select few campus environments. The combination of Ethernet and IP is now the undisputed king of the LAN and, with 1000Mbps gigabit speeds available, increasingly, of the campus. Vendors and industry observers are now even pushing the idea of having enterprises run their time-sensitive, mission-critical voice traffic over their Ethernet/IP networks, thus eliminating the management and expense involved in maintaining a separate network for voice.
Running data over a not completely reliable network isn’t a major concern. After all, if a packet is dropped in a data transmission, it can be resent. But running live voice conversations over unreliable networks is another matter. If voice packets are dropped, a conversation can quickly become impossible to maintain. So the question network managers need to ask themselves is, “Are Ethernet/IP networks reliable enough to support voice traffic?” While industry observers feel the answer for most enterprises is, “Not yet,” a few early adopters are putting their data networks to the test and running voice over IP.