Sunday |
|
|
| Observation Post |
Q&A with Ralph Santitoro, Director of Network Architecture, Nortel Networks, and co-chair, MEF Technical Marketing Committee.For Member Listing - see end of document. Introduction The Metro Ethernet Forum (MEF), established in 2001 and based in Irvine, California, is a non-profit industry organisation formed to promote the adoption of Ethernet technology in metro and wide-area networks. The organisation is composed of more than sixty companies active in the metro Ethernet sector worldwide, principally systems vendors, service providers and test equipment suppliers. The MEF seeks to accelerate the deployment of metro Ethernet technology and services through the production of technical specifications and marketing activities directed toward communication service providers and enterprise companies. Could you give an overview of the Metro Ethernet Forum and its activities? The MEF (Metro Ethernet Forum) was formed in 2001 with the mission of accelerating the adoption of Ethernet technology and services, which is taken to mean the use of Ethernet as a networking and service delivery technology for metro and wide area connectivity - basically all applications of Ethernet outside of the LAN in the building. The aim is to make Ethernet as ubiquitous as Frame Relay in wide-area networks. This fundamental aim has not changed since the Forum was launched. To further this mission, the MEF divides its efforts between technical and marketing activities. On the technical side, the Forum's technical committee develops specifications to address topics that the standards bodies do not. As an illustration, the IEEE (Institute of Electrical and Electronics Engineers) develops the Ethernet protocol and lower level transport technologies, while the ITU-T (International Telecommunications Union - Telecom standardisation sector) focuses on issues relating to carrier networks and how these are built out; neither body focuses on the actual services that can be delivered to subscribers. To fill this gap, the MEF defines Ethernet services and how these are delivered to the enterprise, small business or end-user. During 2004 the technical committee completed eight new technical specifications - this compares to the one issued in 2003. Now contrast this situation to Frame Relay where there is an established terminology that both the carrier and its customers understand. Such a terminology did not exist for metro Ethernet services until the MEF developed standardised definitions. This is an invaluable service for the industry as it means that an end-user can now talk to a carrier, and a carrier to an equipment vendor, using common terms that are universally defined and understood. The MEF's role could be compared to that of the Frame Relay Forum, which developed implementation agreements that pulled together standards from organisations such as the ITU-T to produce specifications that addressed various market requirements. The MEF's marketing committee has the role of educating the marketplace with regards to metro Ethernet networks and services. The committee's activities include publicising the service definitions produced by the technical committee and, more broadly, promoting the fact that carrier-class services can be delivered over metro Ethernet networks today. The latter activity is designed to combat the perception - widely held until recently - that a metro Ethernet network can support only low value, best-effort services and is simply not suitable for delivering high value services with carrier-class reliability and quality of service. Although this view is inaccurate, it was an assumption made by many IT managers and carriers. It was also reinforced by the media, so the MEF marketing committee has focused on addressing this industry misconception by publicising and demonstrating the fact that metro Ethernet is a mature technology and can match the reliability provided by private-line, ATM and Frame Relay services. Has the role of the MEF changed since its launch? In 2001, when the MEF started out, the marketplace knew very little about metro Ethernet technology so efforts were directed towards publicising fundamental features and capabilities. Today the market is generally aware of the technology's capabilities and benefits - so initiatives now address much more detailed questions. However, the principal goal of the Forum - accelerating the adoption of metro Ethernet networks and services - has remained constant. Activities designed to achieve this goal have changed over time to reflect the changing situation in the market. For example, in early deployments, metro Ethernet technology was typically used to deliver high-speed Internet access or Layer 2 VPN (Virtual Private Network) services; today, in addition to these, the technology is being employed to backhaul traffic in support of access networks such as DSL (Digital Subscriber Line) and packet cable. During 2004 the ability of metro Ethernet to deliver triple play - voice, data and video - services was a hot topic with deployments in various global markets really starting to take off. How is the Forum funded? The organisation is wholly funded through its membership fees. The budget raised from membership fees is carefully managed and dictates the programs, and the scale of programs, that the MEF can engage in. Is the membership fee dependent upon company size or is there a flat rate? The fee is the same for all members - currently $15,000 per year. Members can participate in the MEF booth at major industry events such as Supercomm and participate in speaking opportunities at numerous global workshops and events sponsored by the MEF. It also gives access to MEF-funded market research reports, including two Ethernet services OpEx (operational expense) studies that the MEF commissioned, each of which could cost up to $50,000 if commissioned independently. The membership fee offers great value. MEF membership can be particularly helpful to companies with limited marketing funds such as start-ups trying to tightly manage their cash flow and even large companies expanding or augmenting their product portfolio to more broadly support metro Ethernet networks and services. Just consider the cost to acquire, setup and manage booth space at events like Supercomm. For a member company to have a significant presence at Supercomm, as the MEF does, it would cost well over $150,000. That's value! How is the budget divided between marketing and technical programs? The budget for the technical committee is fairly constant as it primarily covers the cost of the quarterly meetings - this program is the same each year. The remainder of the budget is devoted to marketing activities such as event sponsorship, web site development, educational workshops and commissioning of market research reports. Does the MEF evaluate members - for example, if a vendor applied for membership would its products be assessed prior to awarding membership? There is an informal evaluation of what a potential member company does - basically to ensure that it is associated with the metro Ethernet sector. No company fulfilling this criterion would be rejected. The membership consists of networking equipment vendors, service providers, test equipment vendors, component vendors, application developers, system integrators and test houses. Do some members exert more influence than others within the Forum due simply to their size? Companies pursue their agendas in the MEF through participation. The size of the company is only significant if the company elects to drive a particular initiative. Smaller member companies have been able to drive MEF technical committee work just as effectively as larger member companies. The development of technical specifications is wholly contribution driven. Once completed, these must be approved by a membership majority vote where each member company, regardless of size, casts only 1 vote. If one company is pushing aspects of a specification that the majority of other members do not want, it can be voted down during interim approval stages as the specifications are being developed. This way, the relevant sections will have to get modified or removed before final approval. Is the MEF an international organisation and, if so, how does it cater to service providers outside of North America? The MEF is clearly an international organisation given the global usage of metro Ethernet technology. During the first two years of its existence, the focus was predominantly on the North American market. But even at that stage, it maintained an international presence. Approximately half of the MEF's current membership is based outside of North America, although most companies have a global presence with a headquarters and regional or local sales offices in North America even though other functions, such as R&D and manufacturing, may be located elsewhere. Around 30% of member companies are based in EMEA (Europe, Middle East and Africa), including such carriers as BT and France Telecom, and roughly 15% in the Asia-Pacific region. It should be noted that during 2004, the MEF focused a significant portion of its marketing and PR activities in the Europe and Asia Pacific markets. Companies from different regions will have different motives for participating in the MEF, based on the conditions in the local or primary market. Looking at the metro Ethernet market by region, Asia Pacific has been the most active in terms of the extent of deployments, concentrated in South Korea and Japan; Europe is second and North America third. Eastern Europe, in particular, is well poised to become a major market in the near future. Considering the relatively low proportion of members from the Asia Pacific region, I believe that there are a couple of factors involved here. One is simply the travel cost and time required to attend MEF meetings, most of which have been held in North America. Another factor could be the difficulty in gaining visas, particularly in China, to attend meetings in the U.S. The U.S. government has recently removed some of the restrictions for business travellers and this should help. Finally, Asia-Pacific companies may not feel a need at this time to participate in the MEF to accelerate metro Ethernet adoption and deployment since they already agree metro Ethernet is the right technology - and have demonstrated it through their aggressive deployments. Have any companies left the organisation? Yes, a few. The most common reasons for this are one member company acquiring another or a member company going out of business. No company has left for anything other than purely business reasons. Are there companies active in the metro Ethernet sector that you feel should be members but are not? Yes. The MEF currently has 65 members and there are clearly more than 65 companies with some type of involvement in the metro Ethernet marketplace. A company may simply not be aware of the MEF. A company's metro Ethernet products might focus on areas outside of the scope or current focus of the MEF's technical or marketing committee - for example, an Ethernet fibre optic component supplier. Outreach programs are conducted to attract new members and during the second half of 2004, five companies joined the organisation. How does the MEF go about marketing metro Ethernet technology, and does it target both the service provider and enterprise sectors? The MEF markets metro Ethernet technology via several means: a highly comprehensive web site, publication of educational white papers, organising workshops at major industry events, conducting interoperability demonstrations, regular media and analyst briefings, and commissioning studies focusing on the reduced operational costs associated with metro Ethernet versus competing legacy technologies. The MEF conducts workshops, normally lasting for a half or full day, at conferences in Europe, Asia and North America - including Supercomm. Workshops are usually tailored for a specific event, based on the focus of the event - for example, carrier or enterprise and technical or business. The workshops encompass topics such as success stories, business cases, services, technology and network architecture. Such events attract decision makers from both the service provider and enterprise sectors. Regarding Supercomm, the MEF secures a large booth space at the conference each year, runs workshops and conducts a major demo - with members demonstrating capabilities such as quality of service (QoS), network resiliency and service interoperability. Each year, the MEF has conducted a number of workshops at metro Ethernet-focused global events. The most recent one was at an IIR Telecoms event in Barcelona in November 2004. The MEF is recognized as the organization with the broadest metro Ethernet expertise available in the industry and has supplied highly knowledgeable speakers providing compelling content. Event organizers welcome the MEF because they realize its value and our workshops are well attended. The MEF has commissioned market studies to illustrate the cost benefits offered with metro Ethernet technology versus ATM, Frame Relay and private-line networks. The research company Point East Research carried out two OpEx studies for the Forum based on detailed interviews with service providers, one covers North America and Europe, the second the Asia-Pacific region. These studies cover all expenditures associated with the deployment and operation of networks based on the different technologies, including labour costs. A total of 54 carriers were interviewed in these comprehensive studies. MEF members have access to these 50 plus page reports and also have access to an OpEx modelling tool developed using data from the study to perform "what if" analyses of OpEx costs by modifying parameters such as local labour rates. In support of these activities in early January 2005 the Forum launched a web portal tailored towards service providers providing comprehensive information on metro Ethernet technology, products and services - including a database of metro Ethernet solution suppliers. This complements a portal launched in 2004 whose content is tailored toward enterprise subscribers. How difficult is it to get your message across without promoting solutions from specific vendors or, more specifically, is a major company, for example Nortel, likely to join the MEF simply to increase market visibility for its own products? Taking Nortel as an example, the company has been at the forefront of development and deployment of metro Ethernet solutions for at least five years - since the early days. Five years ago there wasn't really a metro Ethernet market of any significance. Nortel, as a founding member, views the MEF and participation in it, as a very effective means to accelerate take-up of the technology. This, in turn, will increase sales of its Carrier Ethernet solutions. However, as many of the major equipment vendors, service providers and test equipment vendors active in the metro Ethernet space are MEF members, the organisation is able to promote solutions in an impartial way - it is able to combine the strengths of the individual member companies to promote the underlying technology rather than specific products. As the market grows, all members benefit. As I stated earlier, the goal of the MEF is to increase take-up of metro Ethernet technology and services, which is obviously in the collective interest of its member companies. If the Forum began to favour the solutions of one company, that company's competitors would object and the organisation could rapidly lose members and credibility in the market. Does the MEF quantify what impact its activities have on the market? I track this in several ways. One indicator I track is the use of MEF terminology in the industry, particularly by service providers when defining requirements in RFPs (Request For Proposals) submitted to equipment vendors. The use of MEF-defined terminology in RFPs is increasing steadily. It is also appearing in equipment vendor product and marketing literature, major industry analyst reports, global media print and online publications, industry webinars, industry white papers and countless presentations given at major industry events. Beyond this informal measure, the MEF website is well used, receiving between 250,000 and 450,000 hits per month. Website visitors are tracked to see who is using the site and which documents are being viewed. Since it was first published, eighteen months ago, the MEF's Ethernet services primer on the site titled Metro Ethernet Services - A Technical Overview has been the most popular. It was also one of the top 10 white papers downloaded from the Light Reading web site for 12 months in a row - Light Reading told the MEF that they had never before had a paper downloaded as much as this one. This is one item that demonstrates the interest in metro Ethernet and additional outreach is still needed to educate and inform the marketplace. Can the rollout of metro Ethernet technology be justified based on OpEx alone, ignoring the additional capabilities it offers? Yes! If a service provider replaced its existing legacy infrastructure with a metro Ethernet solution and simply continued to offer comparable metro Ethernet services - ignoring any new services made possible by the metro Ethernet infrastructure - the reduced OpEx would still justify the transition. The OpEx studies that the MEF commissioned indicate a cost saving of about 21% through use of metro Ethernet as the network infrastructure technology compared with Frame Relay, ATM or private-line. Interestingly enough, this 21% OpEx savings was consistent across the North American, European and Asia Pacific service providers surveyed - even when factoring in different local labour rates. Regarding equipment costs, a metro Ethernet solution will always cost less than a comparable ATM, Frame Relay or SONET/SDH system due to the sheer volume of Ethernet technology shipped. Hundreds of millions of Ethernet ports are being shipped each year and there is no end in sight to this growth. This makes the volume of Ethernet ports shipped each year several orders of magnitude higher than any other LAN, MAN or WAN technology. Are there any changes you would like to see in structure of the metro Ethernet industry as a whole? If asked that question at the beginning of 2004, I would have cited the lack of carriers actively marketing their metro Ethernet services. However, by the end of 2004 the carriers generally changed their outlook and began to aggressively market Ethernet services and adopt the technology. The enterprise demand is there. However, Ethernet services are often not available at all of the locations where they are needed. This is a critical item that must be addressed. In 2003, service providers would say that they were prepared, in principal, to rollout metro Ethernet services or technology but were seriously concerned that they would cannibalise their highly profitable Frame Relay and private-line business. Carriers are now not only deploying the technology but also actively publicising their services - it is no longer necessary to search extensively on their websites to find limited, if any, information about their metro Ethernet offerings. They are now listed prominently beside other data and transport service offerings. Today the carriers' position is that they have embraced metro Ethernet and expect it to eventually become the primary technology for delivering many types of services over a converged network infrastructure. The carriers will continue to offer Frame Relay and ATM services for several reasons. There is still demand, they are currently available in more markets than metro Ethernet and enterprise IT managers simply won't make widespread upgrades to their existing MAN and WAN infrastructures without a compelling business case. This change in carrier attitude regarding metro Ethernet is largely in response to demand from enterprise customers, although hastened by the threat from cable operators that will gladly deliver metro Ethernet services to enterprise customers who request them as a way to take business away from their carrier competitors. Finally, Ethernet services are also being delivered via non-traditional operators such as municipalities who are building metro Ethernet networks to attract new businesses and offer residential broadband Internet services in areas not sufficiently served by the carriers. Do you find it frustrating that the MEF does not address end-users directly, or that end-users are generally uninterested in the technology underlying the services they receive? The MEF's focus is metro Ethernet technology and what it enables; it therefore targets the service provider and enterprise sectors. Residential subscribers really have no interest in how they get the services they want as long as they get them at a reasonable cost, at the quality they have come to expect. I would say that the marketing of competing cable and DSL services has probably confused more people than it has helped. Having said that, metro Ethernet technology does enable the provisioning of lower-cost residential services, and the MEF's promotional activities highlight this point. The organisation currently has no explicit mandate to target the consumer market directly. However, we can very effectively address it indirectly through promoting metro Ethernet as the optimal network infrastructure technology to deliver triple play residential services - using Ethernet or other first mile technologies such as DSL and cable. Does the MEF have a standpoint with regards to the lack of convergence between computing and communications to date, despite the apparent logic of such a development? Convergence has been happening for quite some time and the rate of convergence has been driven by two major factors - bandwidth and cost. As bandwidth increases, new applications and services are possible - provided that the cost is affordable. Take your phone line as an example. Twenty-five years ago, you only used it to make phone calls. Twenty years ago you used it to send faxes via a fax machine or use a 1.2 Kbit/s modem to dial up to a mainframe. Five years ago, you used a 56 Kbit/s modem to connect to the Internet. Modems running at 56 Kbit/s enabled new converged applications such as streaming audio and video but the quality was poor due to severe limitations in bandwidth. This was the extent of convergence over the narrowband analogue telephone line. Then broadband Internet connections, using DSL and cable, arrived and provided 20 times the bandwidth of the phone line, enabling VoIP and decent quality streaming audio and video for webcasts. However, even with DSL and cable broadband, the subscriber has insufficient bandwidth. This limits the deployment of video - television, movies, etc. - which are the most bandwidth intensive residential applications over a common network infrastructure. Offering video and television services is not yet practical because current DSL and cable technologies offer less than a few Mbit/s of potential bandwidth - no guarantees, only best effort service. The ITU-T and DSL Forum are developing standards for higher bandwidth variants of DSL that work over the same copper phone lines entering the residence. Coined ADSL2+, this DSL technology enables up to 24 Mbit/s up to 3000 feet from the customer premise to the DSLAM but the speed drops off as the distance increases. With highly compressed video content, ADSL2+ enables carriers to offer triple play services over that same copper phone line on which you only made phone calls twenty-five years ago. As you can see, more convergence occurs as more bandwidth becomes available, provided that the bandwidth is affordable - but 24 Mbit/s that ADSL2+ provides is insufficient to really eliminate the bandwidth discontinuity between LANs and MANs/WANs. ADSL2+ is only a near term solution. The evolution of the MAN/WAN is clearly based on metro Ethernet and 100 Mbit/s connections, at a minimum - which by the way cost the same as 10 Mbit/s today. With such 'ultra-broadband' connections, the delivery of any variant of video service - HDTV (High Definition TV), SDTV (Standard Definition TV), DVD, and VoD (Video on Demand) - using any compression scheme, can readily be achieved in addition to the less bandwidth intensive voice and Internet data services. With residential services, convergence basically comes down to how much bandwidth can be delivered at a cost the subscriber is prepared to pay. Metro Ethernet technology clearly provides the highest bandwidth at the lowest cost so once the infrastructure is in place, many revenue generating services can be added through simple service provisioning. In the business world convergence is really about sharing information. Videoconferencing is touted as a key application for business, but I see it as ancillary to the real goal of communicating information - it may be 'nice to have' but ultimately does not add much to the information sharing process. The trend now is for multimedia conferencing, where it is possible to have a voice conversation using an Internet connected appliance - laptop, PDA, etc. - in a way unlike using a telephone because you can view, share, annotate live presentations and communicate via instant messaging. I view such capabilities as a more useful form of convergence. Would you accept as one definition of convergence the ability of a user to access any information they require, whatever the format - whether voice, data or video, via any device, wherever they happen to be? Yes, I see this as an aspect of convergence, together with the delivery of personalised information. Looking at convergence from the service provider's perspective, it enables cost savings - for example through deploying VoIP (Voice over IP), where voice and data services are delivered over the same packet network infrastructure. Aside from the economic benefits, convergence also makes new applications possible - VoIP technology enables, for example, a user to listen to voice-mail messages over the Internet. VoIP also eliminates the concept of local or long distance calls. With VoIP, a company can add a voice button to its website which directs the visitor to a call centre agent somewhere else in the world - so a company can cost effectively provide 24 x 7 customer service. Such applications were simply not possible or practical without a converged network. Do you believe that Ethernet is set to be a global standard in the long term? Ethernet is an evolving technology that is still being developed thirty years after Bob Metcalfe invented it (Bob is a MEF advisory director). The original Ethernet technology that Bob developed in the 1970s is very different from that being deployed today. However, the Ethernet brand has remained and it has come to signify simple, low-cost technology, although until recently it was also perceived only as a local, in-building, networking technology - a view that the MEF is charged with dispelling. For the long term, I believe that the only question is regarding whether the Ethernet brand name is retained; that issue aside, I am convinced that the technology will continue to evolve and is poised to become the predominant networking technology for metro and wide-area services. Would it be true to say that as adoption of metro Ethernet progresses in the market, the MEF's role will eventually become redundant? This is certainly true. The MEF's mission is to accelerate the deployment of Ethernet technology in metro and wide-area networks, with the ultimate objective that the technology becomes ubiquitous. So, for example, if in five years time metro Ethernet services are outpacing ATM, Frame Relay and private-line in the market, on top of a substantial penetration, then we have achieved our objective. The indication that this point had been reached would be declining membership, due to healthy equipment sales and service take-up and a general market awareness and acceptance of the technology. At such time the MEF's raison d'être would cease to exist. This is the life cycle of an industry forum. What challenges do you expect the MEF to face in the future? Over the next twelve months we aim to build on achievements to date, through the development of new promotional initiatives and new technical specifications. We have commissioned a new OpEx study to be published in 2005. This one focuses on the economic benefits of metro Ethernet for enterprise companies as opposed to service providers as in the previous studies. We will have more major initiatives in 2005 - stay tuned for upcoming announcements. Do you see any limitations to Ethernet as a technology, particularly with respect to grid computing and delivery of on-demand services? The technology does have limitations, which is why work is continuing in the standards bodies to address future requirements. Development work basically serves to optimise the technology for new applications. Currently, development work is focused on scaling the technology for deployment in large-scale metro Ethernet networks, where millions of subscriber services need to be supported, as is the case with triple play services. To this end, the IEEE has recently formed the 802.1ah Working Group. This Group is tasked with addressing issues relating to security and scalability - particularly with respect to support of large-scale metro Ethernet networks. Service providers are rolling out triple play and Layer 2 VPN services based on metro Ethernet technologies but the size of network and number of users that can be supported is very limited. In effect, a large-scale network is deployed via a series of small adjoining networks. Nortel has implemented a pre-802.1ah technology, known in the industry as 'MAC-in-MAC' which addresses these issues and the lessons learned from live deployments of MAC-in-MAC will undoubtedly help drive 802.1ah standardization work going forward. Another area for development are the OA&M (Operations, Administration and Management) features. Here we can improve the operational efficiency of Ethernet networks. Regarding grid - or distributed - computing, Ethernet is, in fact, one of the few technologies ideally suited to this application. Companies already operate distributed computer networks inside the building, using 100 Mbit/s, Gigabit Ethernet (GBE) or 10-GBE connections. However, extending a gigabit connection outside of the building over a MAN or WAN using legacy technologies is an expensive proposition, so metro Ethernet technology makes this a practical undertaking. Metro Ethernet is helping to make the terms LAN, WAN and MAN obsolete or meaningless. Subscribers will simply have an Ethernet network that will connect to everything - there will no longer be a need for qualifiers such as Local Area, Metro Area or Wide Area. The ability to deploy very high bandwidth connections over a wide area opens up numerous possibilities in the business, scientific and consumer sectors. As an example, with triple play services, a popular feature currently is a personal video recorder that uses a hard-drive to store programmes. Using a metro Ethernet network, this storage could be remote, accessible over the network and the feature sold as a service. Remote storage for primary and backup data is another such example. This is required for business continuity or disaster recovery. With low latency, highly available, high bandwidth metro Ethernet connections to remote storage, computing devices no longer need local storage or may only need some minimal amount. One can extend this further to applications where software applications no longer need to be run or stored locally but can be run remotely so the networked appliance only needs I/O devices - a keyboard, mouse, or touch screen, a video display and a network connection. When the day arrives where Ethernet networks are ubiquitous, Sun Microsystems slogan "The Network is the Computer" can be fully realized.
 |
||||||||||||
