NetWorld+Interop 2003 Tokyo takes place at the Japan Convention Center (Makuhari Messe)
We spoke to Suguru Yamaguchi about the 10th year of N+I.

Suguru Yamaguchi, Ph.D.
Professor, Graduate School of Information Science
Nara Institute of Science and Technology

Question: Could you give us an overview of the last 10 years?
Every year, Interop features a ShowNet Tour booth, which takes visitors on a guided tour of the various exhibits. The tour always covers the most outstanding exhibits of that particular year, but this year there's a table set up with a PowerPoint show presenting a retrospective on the outstanding exhibits of the past 10 years.

Looking back to the first NetWorld+Interop in 1994, we find 235 exhibitors networked over two 1.5 megabit lines, a 10-megabit Ethernet network and a FDDI (pronounced "fiddy") backbone. Moving forward 10 years to the present, we see a venue networked over a 10 gigabit backbone, with each exhibitor booth linked at up to 10 Gbps, with fallback down to 1 Gbps or 100 Mbps, and streaming video, audio and all kinds of applications. This is a completely different world than the one of 10 years ago-it has been transformed into a place where networks are really pipelines, and where computers, which once served as individual workhorses, have now become team players. In today's world, interconnected computers operate in a variety of configurations to provide a range of services, and networks link with computers that no longer even look like computers. Looking back over the last ten years, I think the most interesting thing is that we predicted these capabilities all along, thinking "this will be the shape of things to come, this will be our future." Now that future has arrived, and we are already taking these capabilities for granted.

Yes, yes, that's right, 1.5 Mbps seemed so robust, and, from the tables, we know that booths were linked via a 10 Mbps Ethernet network. In contrast, about ten of the booths today have 10 Gbps access, and a large number are connected at 1 Gbps. Whether that capacity is being fully utilized or not is a complex issue, but exhibitors who are doing demos are streaming video, and some of them are beginning to use the full capacity available. Some exhibitors now say that 10 Mbps is simply not enough capacity, and I'm sure they feel isolated. Looking at the charts, it seems like those voices have increased sharply in number over the past three or so years.

It's common wisdom in the communications industry that computer performance increases exponentially, while performance enhancements in communications technologies progress in a linear fashion. However, I find it fascinating that communications technologies over the past three years seem to be advancing exponentially rather than linearly, and that a technology revolution is truly unfolding right before our very eyes.

Question: Do you think that the use of IP technology has expanded?
I often hear it said that IP has started to cannibalize other technologies. Well, if that's the case, it begs the question, "Is IP is here to stay?" To be honest, it's really hard to say. There was a time when all the latest advances in computer science seemed to be in Unix. And even if you disagree, and think that Unix was just the platform, in fact Unix seemed to be everywhere at the end of the 1980s.

But since then, we have seen new platforms emerge, for example Windows and the MacOS, which have extended their reach to consumers. At the same time, there has been a continuing interchange of technology, and the next platform is beginning to come into view. Its like a spiral-although Windows is really dominant right now, the next platform is beginning to emerge. And the spiral is not spinning in a finite space-actually, it's an ever-widening circle, that's what I mean when I say a spiral. And that's my sense of where things are going.

At the very beginning, networks made up of IBM's old mainframes were, in some sense, the standard. From there, other vendors introduced network architectures, such as AppleTalk and DECnet, after which came IP, OSInet, and a host of other network technologies. And now things seem to have converged on IP technology. In the future, I believe that another technology will grow out of IP, followed by yet another interchange of knowledge that will again give rise to another new technology.

That's why, when people say that VoIP technology was swallowed up, or this or that technology was swallowed up, I would say that these technologies were not swallowed up, but rather that they finished their work. When IP no longer fits the bill, that is, when it no longer suffices technologically, or when new capabilities are needed, then IP technology will probably change, and if the pace of that change is too slow, then a new challenger will likely come to the fore. I think the fact that new technologies appear on the scene to take up new challenges is a healthy thing. I really feel it's a positive thing.

Question: Is this reflected at N+I?
Ten years ago, N+I had a multivendor, multiprotocol environment that ran AppleTalk, NetWare and IP. Over the years, these network technologies eventually all converged to form present day IP technology. In contrast, network applications grew more diverse, applications that include, for example, video, game, and music streaming, storage systems, and IPv6. So while the technology layer has consolidated on IP, applications have broadened, and rather than think that IP has a fixed role to play, or that everything will become fixed around IP, I believe there has been a sharp increase in the requirements posed by these applications. And when someone comes up with something that IP can't handle, that's when the next new technology will likely arise. If, on the other hand, IP can cover it well, then development will likely continue moving smoothly along the current path. This has become evident at Interop as well, where we can clearly see that network technology is still in the consolidating phase, while the application layer is growing by leaps and bounds.

Question: So, with the consolidation of IP technology, will applications have a freer reign?
Yes, exactly. The requirements of these applications are currently being satisfied, you can say, by IP technology. At the point when IP technology no longer meets those requirements, then a new technology will emerge. I think that v6 technology has made some fairly significant inroads. In a way, IPv4 and IPv6 are both IP technologies, but v6 just may be the new, emerging technology to take on upcoming challenges, and we might even be at the point where the work to make the changeover is beginning. Hmm,--- this is certainly one way of viewing the situation. So from a historical perspective, or from the common perception of trends in technology-which we can't really talk about except in hindsight-change is the norm, and, actually, its impetus can come from something quite trivial or insignificant. There are in fact numerous cases of changeovers that were instigated by the simple inability of prevailing technology to serve some need, and I feel we need to stay alert to the fact that this starts from very subtle, simple shifts.

Interop is a venue that offers a glimpse of this-a glimpse into future possibilities; we may not notice it when we see it, but we later realize "Oh yeah, that was it," even though we may not have fully grasped what we saw at the time. Of course, the ability to see it depends on each individual visitor, but of course the idea is to offer a deeper understanding to all.

Question: What is your perspective as a WIDE Project participant?
When Interop was first held, the live demonstration network used at the venue known as ShowNet was initially configured on a network that was created in the U.S., brought to Japan and operated jointly. Along the way, Interop switched to a Japanese network built using Japanese technology, and although it also uses U.S. technology, the Japanese side is taking a leadership role.

Interop serves a variety of functions-it provides the opportunity to track interoperability and check out new technologies. It also enables engineers to establish a community, and lets young engineers learn while gaining on-the-job training experience.

Considering this, it appears that Interop is not all that different from WIDE, although here we use a variety of products, which makes it somewhat different. As a venue for on-the-job training, however, Interop provides engineers the opportunity to get a feel for new technologies. This makes it extremely compatible with the WIDE project, in that it also serves as a place where engineers can come together as a community and collaborate, receive training, verify the interoperability of systems and check out various technologies. Realizing that, Nakamura Sensei and I decided ten years ago to have the WIDE Project collaborate on various aspects of N+I, thereby creating opportunities to see new technologies, to coordinate demonstrations, and to use the event as an a venue for on-the-job training, where students and young researchers could participate in the operation of ShowNet.

In that vein, Interop provides the opportunity for students and researchers to get used to working on site, to experience the difficulties involved in introducing new technologies to a site, and to understand that systems that worked well under isolated conditions in a lab may not function properly in a real operating environment-in other words, to understand the distances involved in bringing products to market. I see a strong connection between Interop and the WIDE Project in that it offers researchers a connection to the real world, and creates engineers who can function comfortably in a real-world environment. Our participation in N+I also allows us to give our students the chance to work together in such an environment, and to provide young researchers with the opportunity to try out new technologies.

Question: What is the relationship between WIDE and N+I?
There are two components to the relationship. On the one hand, there are the areas where WIDE has made further advances, and Interop allows us to see how these advances are being realized-what form they are actually taking. On the other hand, there are tools being developed by users who know nothing of WIDE-in the area of network operation, for example-tools that are of practical use to operators in the field but that wouldn't likely be utilized in a laboratory setting, and Interop provides us with the opportunity to see them. So you can say that there is both a looking back aspect and a looking forward aspect to the relationship. In other words, Interop gives us a bird's-eye view of our progress in technologies where we have moved ahead, and allows us to investigate those we haven't really used ourselves. So I think that both of these components are important in the relationship, but the most important fact, the key word, is real-Interop is a real place where we can try out new things on a real network that is in real operation. In contrast, WIDE's perspective is to discern what the infrastructure of the next networking environment will be, or what communications technology can society next use. These perspectives are not at all at odds with each other, it's just a matter of approach, and in this regard I believe WIDE and Interop have a complementary relationship. This makes Interop a venue of great interest to the WIDE project. Reciprocally, WIDE offers Interop access to new technologies-this time, for example, we have deployed a digital video transmission system-and bringing these opportunities to each other speaks to our complementary relationship.

Question: What about the future of N+I?
Looking at the developmental stages of this thing we call the Internet, we see a period of technological development up to 1995, and a period of quantitative expansion in the seven or eight years since then, through to the present. In short, until now we have been feeding on the knowledge accumulated in the period up to 1995 to significantly extend our capability. Soon, we will need another store of knowledge-not that I'm saying that we haven't been gathering knowledge, of course we have-but I believe that we are now entering into a period of qualitative change. And today, if you look at the showroom floor, you will see new concepts emerging, with several products being demonstrated in ways that they haven't been before, or displayed in novel ways. And while these are still in the minority, I think we are now at the stage where the next breakthrough will be made. I have a feeling that we will see a variety of things change in the next two to three years. The impetus behind this includes, for example, 10 Gb Ethernets, super high-capacity international broadband lines, and the emergence of applications that require speeds of 100 and even 200 Mbps. And these things will have a strong influence on infrastructure technology, network device technology and network operation technology, as the solutions that served us until now no longer suffice. At this juncture, we are passing through the first stage of qualitative change, which I think is the most interesting. I think it has begun, and I think that in the time period from this year, 2003, through 2004 or 2005, we will see a wave of new technologies emerge. While this year is interesting, this will really have begun to take hold by 2004. And so my message for 2003 is that you really must come to Interop and see it for yourself.

  • NSPIXPリンク
  • SOIリンク
  • AIIIリンク
WIDE Award