While floating around last week's Supercomm trade show, one theme became clear. Ethernet is once again very much on the mind of metro equipment manufacturers.
From large established players to small startups, everyone in the metro space jumped on the Ethernet bandwagon in Atlanta. From metro booth to metro booth, the word Ethernet seemed to dominate interest and the conversation.
What's interesting here is how these Ethernet services will be delivered. Just a few years back, there was a whole contingency of chip and board developers focusing on setting up a native Ethernet network in the metro to handle IP traffic. While an interesting concept, native Ethernet networks are gaining little to no attention in the space right now.
Resilient packet ring (RPR) structures have also been the craze of the sector. Packing a Gigabit Ethernet MAC and the ability to optimize the Sonet infrastructure, many have said that RPR is the answer to solving the Ethernet delivery problem.
Despite such claims, all indications show that RPR is also falling behind in this race. While some will contest this point, many agree that for carriers to roll out RPR services, an overlay network is required. An RPR overlay may (and I stress may) be OK for emerging 10-Gb optical structures. But, in the existing 2.5-Gb networks, overlays are seen as expensive and thus are not proving to be an attractive option, leaving RPR in the lurch.
The clear winner is Ethernet-over-Sonet. From large players like Tellabs and Sycamore to smaller startups like Native Networks and White Rock, most players in the metro-networking world spent Supercomm either talking about or showing off ways to map Ethernet frames over existing Sonet infrastructure.
So why is EoS winning the battle. Actually, it wins for quite a few reasons. The first is sparked by the adoption of virtual concatenation. Through virtual concatenation, carriers and equipment manufacturers can carve up Sonet pipes on the STS-1 level, allowing them to pack a host of traffic patterns, such as Gigabit Ethernet into a Sonet pipe.
While Sonet has been a stalwart technology for the telcos, it has also been a bandwidth inefficient approach to networking. Through virtual concatenation, carriers and equipment manufacturers can change this scenario, making the Sonet pipes significantly more efficient.
The adoption of the link capacity adjustment scheme (LCAS) has also proven to be a big winner for the EoS approaches. Through LCAS, carriers and equipment developers can adjust virtual concatenation groups in real-time to deliver additional bandwidth and more. The advantage here is that carriers can additionally change and adjust bandwidth requirements to meet the changing needs of customers, adding flexibility into the equation.
The biggest win for EoS, however, does not lie in the technology. On the contrary, it lies in EoS' ability to keep operational expenses down. By tapping into existing Sonet links, EoS approaches can be implemented without carriers having to make big changes to their element and network management systems. In this tough telecom world, that's a real win for the carriers.
Of course equipment developers and carriers must resolve what technology to use when rolling out 10-Gb metro networks. Many will contend that existing fiber can't handle all this and a new network will be needed. If this is the case, then RPR could yet be an attractive solution for equipment developers and carriers.
But, don't count out EoS in this space as well. While some will argue that new fiber optic cables and networks are required, there's a whole set of people looking for ways to optimize existing fiber for 10-Gb operation. If these efforts succeed, then making the leap to a new network may get pushed out further, once again playing into the hands of EoS.
Robert Keenan is the Editor-in-Chief of CommsDesign.com. When not attending Supercomm, Rob can be reached at rkeenan@cmp.com.
