Allied Telesis: Supplying Building Blocks for Networks
An interview with Anders Swahn and Mark Nakayama of Allied Telesis Allied Telesis, a leading supplier of building blocks for local area networks, specializes in selling LAN components to Japanese systems integrators and end-users. In October, Computing Japan talked with Anders Swahn (vice president, product marketing) of Allied Telesis Inc., and Masanari "Mark" Nakayama (senior vice president, sales & marketing) of Allied Telesis KK, to discover how the company has managed to build up such a commanding lead in a very competitive field.interviewed by Terrie Lloyd
Computing Japan: I understand that Allied Telesis has had a very good year in 1994.
Mark Nakayama: Our 1994 financial year is not yet over, but we are on track for an increase in shipment volume of around 250%, and in shipment value of 150%. In actual revenue, the increase has been from just under ·4 billion last year to ·6 billion this year. And because the LAN market is predicted to grow by 150% in 1995, we are looking for a sales increase of least or better than this amount. In terms of personnel, we have grown from 140 people at the beginning of last year, to 190. Because of the growth, we moved in August 1994 to our new Japan headquarters in Gotanda. Around 100 staff work at the new premises; the others are stationed in Fujisawa (60 people), Osaka (18) , Machida (11) and Nagoya (4). Our main engineering and warehouse centers are in Fujisawa.
CJ: What percentage of Allied Telesis worldwide revenues come from Japan?
Nakayama: About 30%. Our revenues worldwide are very evenly distributed; we make about 1/3 of our sales in the US, 1/3 in Europe, and the remainder here in Japan.
Anders Swahn: The three markets are at different stages in their maturity. The most mature market is the North American market, where network growth is slowing; for example, in 10Mbit adapter cards, it is predicted that in 24 months there will be zero growth. The connectivity rates in business are as high as 60% to 80% of all PCs.
The next most mature market is Europe, which has lower connectivity rates of around 20% to 30%. The youngest market is Japan, with a very low connectivity rate of 10%. We see the best prospects in this fast growing market.
CJ: What would you say are the main things that have made you successful in Japan?
Nakayama: Support. By their very nature, LAN users cannot afford downtime, so support by makers like us is very important. Many companies are entering the LAN market because they can make profits; however, a large number of small players soon pull out again and leave the users without support. Allied Telesis has about 30 technical support staff, including an end-user technical call center with five hotlines. And we have a field service operation which does second-level support for our dealers when they have problems.
CJ: With the high rate of growth you are experiencing, Allied Telesis has moved from being a small company to a medium-size company very quickly. To maintain that growth, are you considering any alliances?
Swahn: In sales, across the board, we are doing a lot more OEM manufacturing. In Europe, for example, we sell to such customers as IBM Europe, Siemens, Alcatel, DEC, and other large firms into ATM. On the technology side, a major alliance for us is FORE Systems, the leader in ATM technology in the USA. The product we're developing with them is LAN emulation software.
Nakayama: We OEM four different products here in Japan. I can't reveal who we do the work for; however, I can say that OEM sales represent about 10% of our turnover. We are not specifically looking at chasing OEM work, but we don't turn down the business if it comes our way.
CJ: What products most profitable for you right now?
Nakayama: There aren't any products that really stand out ó the sales increases were across the board. Transceiver shipments grew as anticipated; our market research tells us that we have a 50% market share in transceivers in Japan. (We make our transceivers in China and produce about 130,000 to 140,000 units monthly for distribution worldwide.) For repeaters, we have about 30% to 35% of the Japanese market. Our share of the hub market is still low at about 20%.
CJ: Who are your major competitors?
Nakayama: It depends on the type of product. For example, for 12-port hubs, of which we have 30% of the market, the competitors are 3Com, HP, and D-Link.
CJ: Controlling costs seems to be a major issue for your product range.
Nakayama: That's true, especially in the lower end hub products. The chip sets are all basically the same, so it is manufacturing cost control and unique marketing that give us differential. We could improve on costs if we can move from the metal case used in the States and Europe (because of strict EMF noise regulations) to a plastic one here. Plastic would give us a big advantage. Another way is to manufacture in low-cost countries, including Thailand, Singapore, and China. When manufacturing in these countries, though, we always take the precaution of producing the same model in at least two places to allow for stoppages and supply shortages. We make 80% of our product volume in Asia and the remaining 20% in Japan and the USA.
CJ: Do you have your own factories in Asia?
Nakayama: No, we use partner companies who subcontract to us. Because of the volumes, they are generally working 100% for us, but we prefer them to be separate entities.
CJ: MIS managers now have a tremendous number of choices in planning for their corporate networking needs. What do you see as the likely mainstream direction of networking over the next few years?
Swahn: We're taking a very focused approach to what markets we pursue. We want to be good at what we do, and not dilute our efforts in too many fields. Allied Telesis has selected technologies that it believes will be successful in the future, and we are spending all our R&D dollars there. Today, the 10Mbit standard is predominant. We believe that ATM will be predominant in the future ó first, because it is scaleable in bandwidth; second, because it works well for integrating voice, video, and data; and third, it reduces the costs of owning a large-scale network because it is easier to manage [than router-based networks].
Our R&D strategy is to help the customer migrate from today's Ethernet networks to the ATM networks of the future. We are not pursuing FDDI or Token Ring development.
CJ: The Gartner Group, IDC, Dataquest, and others have said that the costs for managing client/server networks are actually very high, and not necessarily compellingly favorable versus a traditional mainframe. Do you agree with this, and how are you addressing the problem?
Swahn: We're doing a couple of things. I can't go into details because the products haven't been announced yet, but we're putting together a strategy where we have graphical user interface solutions for many different platforms, so as to reduce network management costs for larger networks. Another thing we are doing is to develop virtual LANs, so as to allow customers to reduce their overall costs of managing networks. Even though the cost of network ownership today is higher than what people thought they would cost originally, it is still much, much cheaper than any mainframe implementation, quite apart from the scaleability offered by client/server technology.
CJ: Not so long ago, Cisco Systems Japan formed a powerful consortium of 13 Japanese big-league networking players to promote the router solution for inter-networking. How are you answering this challenge?
Swahn: You have to look at the inherent advantages that ATM technology brings to the customers. If you look at the US, for example, where a lot of ATM R&D is taking place today, there are multiple ways that people can apply ATM in their networks.
One way you can do it is to use the Cisco/Fusion solution, where the ATM is integrated into the router. Another alternative is to have the routers performing a different function in the networks. For example, they will not be the central point of the network, and the networks will not be "router-centric." I think that users will find, over time, that there is a very attractive, very easy-to-manage alternative to building networks, where you use ATM backbone technology and where it will to some degree reduce the need for routers. Maintaining a router-based network is very manpower intensive, particularly as people move around. I think you'll see that this new technology... I hate to use the term "paradigm shift," but there are new ground rules in the industry, and the market will figure out how this new technology will allow people to rethink how to build networks. It will let them integrate voice, video, and data, reduce the dependence on routing technology, improve security and performance through virtual LANs, and allow employees to move around without having to do lots of software maintenance on the router network. This whole process of how you migrate your network to ATM is something of a learning experience. The Cisco thing is one way of doing it. A large number of independent companies will be doing a little more novel thinking about how they apply ATM on a non-router-centric model, and we are just one of those companies.
CJ: What products are you releasing to help Japanese companies move to ATM?
Swahn: We have a programmable switch that lets you either store-and-forward or cut through. There are two camps of thinking about which method is best, and we have built the capability through software configuration to handle both.
The Cisco view of things is that the router is the center of the universe. However, we think the LAN emulation/ATM camp is evolving. ATM switches are at the center of the network, with switching hubs feeding into these networks. The routers have two functions: one is a LAN-to-WAN function, and the other is a route-server, to route all the virtual LANs in the network. This lets you physically put the router in one location and route between the virtual networks, significantly simplifying your network management and maintenance.
CJ: In what applications does ATM really excel?
Nakayama: A typical application for ATM is for graphics ó routers are really too slow. At Interop in Atlanta last year, for example, 3Com set up an animation display using Ethernet and routers ó but it didn't work that well because of the router speeds. If they had used ATM directly, they could have had perfect animation. But it would cost ·2 million per port for a customer to do this. A better way to do it is to use Switched Ethernet. At a display we did recently, we had a similar setup where we ran animation over 12 different PCs and it worked really well.
CJ: Until now, Allied Telesis has been a vendor of basic building blocks for networks. With more complex solutions on the way, will you be butting heads with the big boys? What is your strategy for competing?
Swahn: It is true we are a building block company, and we're not going to change our strategy. We will not evolve into an end-user solutions provider. We don't take our products and sell directly against Cisco. Instead, we target them at people in the ATM switch business who might be using them as part of an overall integration effort. We don't see Cisco as a competitor; our products can be used in a router-based network as easily as they can be in an ATM switch-based one.
CJ: What do you think about the Synoptics-Wellfleet merger and the fact that Cisco is now supplying router products to Cabletron?
Swahn: Our industry is going through a general consolidation. There are a lot of small companies who are not doing very well; a number of smaller hub companies are going out of business. The larger companies are consolidating because they believe that they need a broad base of solutions, since they are in the business of providing end-user solutions. The Synoptics-Wellfleet merger is a good example of two companies that are trying to solve all the end user's problems with their product range.
We, on the other hand, have a very different and focused strategy. We're a building block supplier, not an end-user solutions provider. We believe that at the end of this industry consolidation, Allied Telesis will come out much stronger. We see these events as a positive thing for us.
Late last year, Allied Telesis, Inc., changed its corporate name to Allied Telesyn International Corp. The ATI Group companies in Europe have changed their names as well, but Allied Telesis KK (Japan) and Allied Telesis Asia, Pte Ltd. (Singapore) will retain their names.
Many companies are entering the LAN market because they can make profits; however, a large number of small players soon pull out again and leave the users without support.
We're taking a very focused approach to what markets we pursue. We want to be good at what we do, and not dilute our efforts in too many fields.
This whole process of how you migrate your network to ATM is something of a learning experience.
We are a building block company, and we're not going to change our strategy. We will not evolve into an end-user solutions provider.
