Metro/Carrier Ethernet Continues Its Conquest

By Richard “Zippy” Grigonis

Metro Ethernet is a fast-growing market for service providers. Metro Ethernet services are often called Carrier Ethernet by providers, because they now can cover the whole wide area too. It provides services primarily to enterprises and can save them a lot of money, especially compared with older technologies such as SONET, frame relay or ATM. (Of course, you can deliver Ethernet over those services too, if for some strange reason you really want to.) Ethernet is being pushed as being simpler, cheaper, faster and of generally higher performance than competing forms of transport. Still, there’s an old saying in the network business: You never get rid of anything. You just depreciate the assets or the customers just tend to buy an additional service. In the case of etro/Carrier Ethernet however, operators are indeed dropping frame relay, SONET and ATM in favor of the new upstart.

Equipment vendors are having a jolly good time with Ethernet, despite the current recession. Take ANDA Networks, which provides carrier-class Ethernet equipment solutions for delivering Metro Ethernet services over fiber, copper, and wireless based access networks worldwide. They allow next-gen fiber-based networks with managed QoS offerings to maximize bandwidth with incremental levels of Ethernet service while lowering operational and capital expenditures. They’re also adept at the rapid deployment of Ethernet service to business end-users on traditional legacy networks and generating new revenue opportunities by extending the reach of Carrier Ethernet service. Finally, they allow Ethernet traffic bundling via intelligent devices situated at multi-tenant buildings or at a service providers’ POP or CO.

ANDA’s Greg Gum, CMO, says, “The impact of the current recession varies, depending on what customer base you’re addressing. The carriers were fairly insulated up until just recently. Indeed, most carriers tend to be fairly well insulated during recessions because of their large capex and recurring revenue base. Given the fact that a lot of consolidation has already taken place, the general trends are that they can hold back from their capex spends and/or measure them relative to what they call success-based deployments, and be able to weather the storm much easier than other companies that have major ebbs and flows in their cash flow. People don’t get rid of their phones. Even in a recession, people tend to use their phones, wired or wireless, even more, since they make calls instead of traveling. So, from a general perspective, the Metro Ethernet industry has been fairly robust.”

“Over the last two quarters of 2008, we had two up quarters, compared to the first half,” says Gum. “Even in recessionary times, we see several trends. One is that the fiber deployments were pretty much solidly earmarked for things in the North American markets and even in Europe where they had committed their fiber spends. They weren’t changing those plans very much, particularly in the business market, for business Ethernet services, which is where we focus. A second trend is that, from a copper-based deployment perspective, we’ve also seen a fairly good ramp-up of folks using Ethernet over traditional circuits, such as T1s and E1s and DS3s and E3s, to provide Ethernet over existing PDH circuits. Part of that trend is also because of several things. For example, if they didn’t have the fiber built out to the customer, they would then use their existing circuits and put Ethernet over the top of those in the interim, rather than try to deploy new fiber. They could do that because these are typically business-space deployments. So the trend during the last half of 2008 is that, even in a recessionary time, the operators could use either the copper solutions that we have if they need to conserve costs, or if they were continuing their fiber build-out, and they were just building off of their fiber spurs, they could then use our fiber products.”

“So, from an ANDA perspective, we actually saw very good third and fourth quarters in 2008,” says Gum. “And going into Q1 2009, we have a pretty good backlog of orders that we’re trying to build out and ship this quarter. The applications that are driving this from a customer perspective are, first, we’ve seen several North American carriers basically ditch frame relay and move to an all-IP network, specifically one based on Ethernet at layer 2. They’re retiring a lot of the old frame relay equipment and in Europe, some of the ATM equipment. This is because of cost and the fact that much of this equipment is about 10 or 15 years old and the operators have taken all of the depreciation of those assets. The second trend is that, as the Verizon FiOS, AT&T U-verse and European fiber ring concentration penetrates from the cities out into the outlying network, operators are now starting to build off spurs to business sites using fiber. We’ve seen the bandwidth usage pick up too, because of several things: one, Ethernet is now de rigueur as a standard for new WAN services. There’s been a pickup in carrier bandwidth, either interconnecting to other carriers, thus bringing on wholesale Ethernet opportunities, as well as bandwidth increases owing to the upgrade in the mobile networks, which has driven much of the bandwidth capacity increases for the network. For example, if you move from 2.5G to 3G or from 3G to LTE, most of the carriers are trying to offer between 20 to 25 megabits per subscriber in a new 4G LTE network to handle streaming applications and mobile TV, which is a considerable amount of bandwidth.”

Beating the Competition to the Punch
Ciena’s Carrier Ethernet Service Switching Family integrates Ethernet services management, a common service-aware operating system, and programmable network devices to aggregate and deliver business and transport Carrier Ethernet services from the edge of the metro/core to the customer. Ciena’s True Carrier Ethernet combines the low cost and high capacity of Ethernet with the reliability, management, and service quality associated with SONET/SDH networks. Ciena also offers their Carrier Ethernet Service Extension Family, so that service providers can now extend revenue-generating Ethernet services to end-customers over fiber or copper access networks, and provide Carrier Ethernet transport for legacy TDM and ATM services for converged voice/data business services and wireless backhaul applications.

David Parks, Director of Product Marketing at Ciena, says, “There are many challenges out there now, even if you’re just looking at the macroeconomic environment. For service providers, regardless of what kind of provider they are – telco, cable MNSO, large or small – they’re all increasingly competing in terms of time-to-market. Related to that is time-to-revenue and time-to-profitability. One reason this is occurring is because Carrier Ethernet services are more widely available than they were a few years ago. Chances are, if you’re a business customer looking to buy a service, you might have two or three possible providers with which you can deal. A couple years ago, even you could even get the service, there might be just one service provider in your area. Things are now good for the end user, but it presents some challenges for the service provider. Providers also need to find new ways to differentiate the services. That’s where we’ve been placing much of our focus these days. We ask ourselves, what can we do with our equipment and solutions to help our carrier customers differentiate their Ethernet services? That’s where ‘service velocity’ also comes into play. The service provider who can install, configure and provision the service in a very fast and efficient manner will be better positioned. That’s where we place much of our development right now – developing software functionality that helps the provider do that.”

“Getting back to the economic climate,” says Parks, “we’ve found the nice thing about Ethernet is that it enables you to do more with less. You can get more bandwidth at a lower cost per bit, and you can get more flexibility in the service. You can extend service to more and more locations at low cost, so in a down economy there’s a silver lining for Ethernet.”

Troubleshooting Metro/Carrier Ethernet
Companies cut back wherever they can, and carriers are now different. Fluke Networks is introducing a new way to help these providers to test and troubleshoot Carrier Ethernet lines. Fluke Networks is known for its solutions for the installation and certification, testing, monitoring and analysis of copper, fiber and wireless networks used by enterprises and carriers. The company's extensive line of Network SuperVision Solutions provide network installers, owners, and maintainers with the ability to highly optimize network performance. Recently they introduced the latest version of their MetroScope analyzer and the ProVision test suite that allows providers to quickly test for throughput, loss, latency and jitter in parallel.

Mark Mullins, Fluke’s Carrier Ethernet Marketing Manager, says, “Many customers think that simply doing a speed test on their Ethernet connection is a good way to verify whether or not they’re getting the performance for which they’re paying. The problem with those speed tests is that you’re testing many things that aren’t really part of the circuit and you’re not measuring a lot of things that might be important to you. If you look at the ITU guidelines for the performance of a circuit, you’ll notice that all sorts of things need to be measured, such as delay, delay variation and loss. Obviously, those quick web-based tests don’t give you those kind of performance measures. That’s why as a customer or carrier you should care about testing – a lot of high performance applications, real-time services such as VoIP or video teleconferencing or even highly interactive things such as online transaction processing, require that you have short delays, short delay variation and you keep packet loss to a minimum. That’s why some of these other measurements have come up and that’s why RFC 2544 testing has made a lot of inroads in the industry. If people do a thorough test on a Carrier Ethernet circuit, they’ll tend to use the RFC 2544 tests, which measures the five main factors: throughput, loss, latency and jitter being the most important ones – to really characterize the performance of a Carrier Ethernet circuit.

“If you’re installing a new circuit, you go to the customer’s site, hook up a tester to each end of the circuit, and then run the RFC 2544 suite of tests and then you’ll really know what the performance of the circuit happens to be,” says Mullins. “You could do this when you’re installing a new circuit, or troubleshooting an existing one. It’s become a standard test. Additionally, we’ve developed a device called a Reflector which allows you to test circuits without having to buy a second fully-featured product. The Reflector costs about $1,000, as opposed to an $8,000 test set. You can put the Reflector at the far end of the circuit and use that for the test. When we debuted the MetroScope about a year ago, that was one of our first price ‘breakthroughs’ to reduce the cost of testing by about 40 percent. That’s turned out to be very popular.”

“Based on our expertise with IP, our MetroScope can show you the traffic levels, protocols, who’s doing the talking on the network,” says Mullins, “and we can actually discover all of the devices that are on the network. We can analyze the VLAN connections. This and the Reflector have been our biggest advantages over the competition. But it has also restricted the size of the market that we can pursue, because, in the Carrier Ethernet world, there are a few technicians or even engineers who can take advantage of all of this ability to test operating networks. There are many more technicians out there who just want to plug it in, run a test, and tell you whether or not the circuit’s good – that basic RFC 2544 type of testing. That’s my lead-in to our new product, the MetroScope LT.”

“Since we’ve found that many technicians don’t have either the need or the experience to use much of this capability to test the operating networks,” says Mullins, “we decided to take it out and make a less expensive version for those kind of technicians that doesn’t have all of those extra features that they don’t really need and/or they don’t really know how to use. That’s what our new MetroScope LT is all about. It still has the ability to measure the network performance, but when it comes to debugging the actually traffic on the network, that’s a small part of the job for most people, we took out the associated pricing and cost and voila, we got what we call the MetroScope LT, which is just a fancy way of saying ‘Lite’. It’s a base model. Between the MetroScope LT and the Reflector, you can actually test a circuit for half the price of the absolutely cheapest thing that’s out there today.”

Meanwhile, over at Soapstone Networks, they develop resource and service control software that automates the service lifecycle for carriers and large enterprises. Soapstone enables customers to evolve their network and service offerings independently by bringing centralized control to transport and separating services from the underlying transport technologies.

Esmeralda Swartz, the Co-Founder of Soapstone Networks, says, “Soapstone provides a resource for a service management framework that works across a multivendor network. In the Carrier Ethernet space, many people talk about it as if it were around for a while. The reality is that only recently with the emergence of some of the OEM functionality, and some of the new hardware that you now have the ability to leverage Ethernet technology in a big way in carrier networks. What that means is that there is a lot of functionality, particularly around the OAM area, of which carriers want to take advantage. One of the challenges, however, is that there are varying degrees of support across all of the equipment suppliers. One of the jobs that we perform is allowing a carrier to in effect build a multivendor network of equipment that will all work together. We are essentially providing that management framework that deals with the interoperability on the data plane and is also able to provide a common OAM capability across all of the supported vendors.”

“We essentially abstract the different vendor implementation differences across some of the supported standards,” says Swartz. “One of the reasons that carriers have difficulty in building out multivendor networks is that they themselves end up having to deal with the vagaries of different vendor implementations. So what they’re really looking for is a way to provide an end-to-end service experience across all of these different domains and across different NEPs. One of the challenges is, how do you manage that end-to-end service? That’s where Soapstone’s solutions makes its mark. Our job is to understand the service from the access network to the metro network and then to the MPLS core network.”

“I would say that another big challenge is that carriers in particular having to pay a lot closer attention to what the intended service behavior actually was, and what the network’s actually doing,” says Swartz. “One of the aspects of that is the ability to do rapid troubleshooting, so they need a system that allows them to view the network in real time, in essence correlating what the service request was, what the network is actually doing and how you respond to issues that occur in the network.”

The Evolution to Ethernet
Redback Networks (an Ericsson company since January 2007) specializes in providing scalable solutions for fixed and mobile carrier networks that can deliver next-gen broadband services such as VoIP, IPTV, on-demand video, and online gaming. Based in San Jose, California, Redback Networks has more than 500 carrier customers worldwide.

Jeff Baher, Senior Director of Product Marketing at Redback, says, “Ericsson acquired us because of their intent to move to an all-IP infrastructure. In particular, Ericsson’s bent is toward mobile communications, and on the mobile side certainly, over a long time period, almost everything in the 4G infrastructure will go all-IP. Through the years, Ericsson had a partnership with Juniper and was doing most of their R&D through basically delivering their software through Juniper platforms. In particular, the GGSN, and other Ericsson products were coming from Juniper and other IP vendors. But Ericsson was taking those and building solutions around those point products. That worked for a number of years, and Ericsson had a Number 1 position in the GGSN or the xGSN market, by developing their own software and delivering that on Juniper platforms. But around 2001 it was pretty clear that these infrastructures and the role that IP was playing, was becoming a lot more significant. Ericsson needed to have a tighter relationship among the IP technologies they were selling through in their solutions, and that’s what drove Ericsson to look at what IP vendos were out there – Cisco, Juniper, Redback – and then determining what vendor they could essentially acquire so they could exercise their own control over the IP technologies in these solutions.”

“Ericsson bought into our SmartEdge technology and platform because, over the years, Redback had generalized their platform to do IP and MPLS routing,” says Baher. “So in addition to the subscriber management, the Redback platform has emerged as a bona fide carrier class IP router along the way. And, for a number of solutions that Ericsson sells, they needed IP routing platforms in their own portfolio. So since the acquisition we’ve worked aggressively to get our SmartEdge platform engineered into as many of the solutions as Ericsson sells as possible, such as mobile packet backbone, mobile backhaul, and various solutions that basically support the evolution of the mobile infrastructure.”

“In that same period of time that we’ve positioned the SmartEdge into these Ericsson solutions,” says Baher, “we also introduced a new platform, the SM 480 Metro-Ethernet Service Transport, which we introduced in June 2008. The motivation behind this was to build a platform that had a stronger focus on Ethernet for metro applications. That meant that there’s a few things that weren’t required in terms of interfaces and the level of IP intelligence. So we were able to harness the Ethernet requirement and cost-optimize the feature set, so we could build a platform to go deep into the metro area to deliver Ethernet. It works and complements the existing SmartEdge, but it allows the carriers to extend their packet footprint farther into the metro area.”

“We would agree that as carriers migrate to a packet infrastructure, the dominant interface that things will eventually converge on will be Ethernet,” says Baher. “From that perspective, Ethernet doesn’t represent necessarily a new platform or a new specific location in the network. It really just refers to the changeover from traditional TDM interfaces or ATM interfaces on switches and routers to Ethernet interfaces. We can see that at peering points which traditionally might have been OC3 or OC12 connections. Now you’ll see Ethernet or Gigabit Ethernet peering. Carriers are looking to adopt Ethernet technologies even within the POP as you connect edge routers to core routers. The interfaces were historically SDH packet-over-SONET interfaces. Those are now becoming Ethernet interfaces.”

Something for Nothing?
Sorrento Networks is a global provider of metro optical access solutions. Sorrento's WDM and ROADM products enable carriers and enterprises to inexpensively add capacity to their existing optical networks without investing in additional fiber cabling, thanks to optimization. Sorrento recently unveiled a new Ethernet-centric product.

Jim Nevelle, CEO of Sorrento Networks, says, “We are a metro WDM equipment provider. We picture this world in three parts: First, the long-haul core infrastructutre which we picture to be the Alcatel-Lucents, Fujitsus, and those fellows. Then you have the metro players, which is us. We play in the world of Transmode, BTI and those folks. And then we also deal with the access portion of the network, which is more of a last-mile CPE environment. In the metro world we have customers such as Cox, Comcast and tw telecom. We deploy overseas in places such as Neos Networks, a tier 2 provider in the U.K., and Deutche Telecom in Germany. We also have deployments in southeast Asia and Australia.”

“So we’re heavily deployed in the metro environment,” says Nevelle. “We’re seeing Metro Ethernet driving the carrier space right now. If it’s not their number one priority, it’s at least up there in the top set, with every one of our carrier customers. Everyone wants to know how to deliver Gigabit Ethernet services and offerings out to their customer base more effectively. We are seeing companies that might not have had the this initiative in January 2008, but now it’s definitely their initiative in January 2009. Even with the interesting economic times we’re seeing today, everyone is still basically putting their resources and time and effort into more and more delivery of Ethernet services, especially into the metro environment.”

“First, Ethernet enjoys ease of use,” sayd Nevelle. “It’s readily accepted and adoptable. People like it. It’s uniform, standardized. And it’s relatively inexpensive compared to competing technologies, so it allows people basically jump on the bandwagon, if you will, and take advantage of the economies of scale of which Ethernet is capable. People now also see with Ethernet that it’s stable. You can actually use Ethernet now as a transport vehicle. It’s no longer being contained within an office building on a LAN. You can now utilize it within a WAN environment, particularly in the metro area, and receive just as good if not better quality of service than you were getting prior to adopting Ethernet.”

“Since we deploy in the metro world, all of our customers ask us how to effectively deliver Ethernet solutions, particularly Gigabit Ethernet solutions out to the customers,” says Nevelle. “Hence the reason why we rolled out our new optical Ethernet transport card. What we’ve done with the OET card is to allow customers to go out and fire up a 10-gig wavelength in a metro environment. The customer can now tap into it anywhere they go. So here’s one of the ‘nuances’ of metro deployments. If you get into the core world and start deploying large WDM infrastructures, you know you’re going to achieve some economies of scale, because you have a backbone infrastructure and you have X number of circuits thrown on there, and you know you’ll get some sort of return on investment, because it’s an aggregation point. The problem is that, when you flip over to the WDM metro world, let’s just say that everybody’s crystal ball isn’t as clear. You know you have a ring, you know you have some capacity in the field somewhere, just as going through a city, but you just don’t where the next circuit going to light up. You don’t known where the salesperson is going to sell the next circuit. That causes issues, because the provider doesn’t know if he should build his infrastructure ahead of time or wait until the actual order appears in hand. What do they do? Ethernet pricing has basically achieved a sort of parity, so now the question becomes how fast you can deliver these circuits. No one has the money to build the circuits and advance and have them ready and waiting. So in the metro world they wanted to find a way to minimize the initial cost but allow them to deliver services faster. And so with our OET card we can light up a 10-gig wavelength circuit all the way around the ring and then allow the customer to tap into that 10-gig of Ethernet bandwidth anytime they want, wherever they need it.”

Everybody Loves Ethernet
It’s true. Long-range versions of Ethernet are slowly displacing older TDM, frame relay and ATM transport. Ethernet has too much going for it (cost, ease-of-use and quick deployment) for any competing form of transport to dominate.

Richard Grigonis is Executive Editor of TMC’s IP Communications Group.

Companies mentioned in this article:

ANDA Networks
www.andanetworks.com

Ciena Corporation
www.ciena.com

Fluke Networks (News - Alert)
www.flukenetworks.com

Redback Networks
www.redback.com

Soapstone Networks (News - Alert)
www.soapstonenetworks.com

Sorrento Networks
www.sorrentonet.com

NGN Magazine Table of Contents