Voices from Clients
Case Studies

Pursuit of new challenges forms the core of IIJ’s values

Ever since its inception, IIJ has been on the leading edge of Japan’s Internet business. The containerized data center that you’ve built was the first of its kind in Japan. How was this project conceived?

• 

  Internet Initiative Japan Inc. (IIJ)
  President (now Chairman & CEO)
  Koichi Suzuki

• 

  Matsue Data Center Park (an image by computer graphics)

Suzuki:：The idea of a containerized data center had been floating around for several years, and it seemed to have many advantages, including lower power consumption, ease of scaling, and natural air cooling. So we embarked on R&D on this concept two years ago in cooperation with an air conditioning company and a shipping container manufacturer. There were a few obstacles, however, including the high cost of land acquisition in Japan and sizeable electricity bills (which could be as high as 60% of the operating cost of a data center). Since air conditioning accounts for a large part of the electricity use, we focused on pulling cool outside air into the facility and finally succeeded in bringing the power usage effectiveness (PUE) to a satisfactory level. That was how this Matsue Data Center Park project got off the ground.

I can see why they say IIJ is farsighted.

Suzuki：Well, I wouldn’t call it farsighted—just adventurous, I guess. (Laughter.) I mean, whenever I get a sense that things will go in a certain direction, I’ll follow that instinct and go all in. Of course it’s hard to tell when such a prediction will come true, and Japanese people are generally wary of trying new things, but IIJ is not like others. The reason why this company exists at all is because we love doing things no one has done before. Admittedly, any new solutions are hard to sell, especially to Japanese customers. American clients readily show interest in new things and want to try them, whereas Japanese clients typically ask, “What other companies are using them?” But where’s the fun in doing business unless we get to do something new? So we built the first containerized data center in Japan and are also pursuing many other new projects.

I gather you also came up with the idea of cloud computing quite a while ago?

Suzuki：Yes, we had a go at cloud computing as early as in the 90s. We thought the time would come when companies no longer needed to keep data processing systems in-house and built one of the earliest data centers. But it was too far ahead of its time to become a commercial success. It’s really difficult to match an innovation with market needs. I tend to push innovations before the time is ripe; I can show you a long list of failed projects. (Laughter.) But I have a feeling that the Matsue Data Center Park will mesh with the customers’ needs. Its commercial outcome still remains to be seen, but technically the project has put us far ahead of our competitors.

It’s remarkable that you had foreseen the arrival of the cloud computing age more than ten years ago.

Suzuki：In my opinion, the change was too slow to come. Once an idea that could shape the future is recognized in the U.S., it takes just a year or so to build a market around it, whereas it often takes as long as six or seven years to do the same in Japan because people here are on the whole overcautious.

IIJ’s technological prowess is widely known. Can you speak about that?

Suzuki：Well, IIJ and NTT Docomo were founded in the same year, and at the time we felt we could easily outgrow NTT Docomo, but it didn’t work out that way. It was partly my fault, because I talked NTT into entering the Internet business. (Laughter.) We even offered them technical assistance to create OCN, which grew into the largest Internet service provider in Japan. NTT Docomo called us the “father of OCN” at the time, but it was meant to be derisive because they robbed us blind in some market segments. The same thing happened with NEC’s BIGLOBE and other Internet service providers: they all thrived by piggybacking on our service package. Still, all of us at IIJ were proud to be pioneers of the Internet world. The PPP dial-up protocol was among the many technologies that IIJ developed. It is an epoch-making technology in the history of the Internet and is recognized as such in overseas tech communities, although it is much underappreciated in Japan. IIJ’s technologies have certainly helped Japan carve a large share of the global Internet market, but whether those technologies equally helped IIJ’s business is questionable.

More data center parks to be built in Japan

Demand for servers is surpassing supply in the aftermath of the Great East Japan Earthquake, is it not?

Suzuki：True, demand for servers is now higher than ever, especially because many data centers are operating below capacity due to the short supply of electricity. At present, the Matsue Data Center Park has only two container modules, but it will be increased to seven by September. If a comparable scaling were to be attempted at an urban data center, it would easily take a few years, which is a big disadvantage. We will continue to expand our operations in Matsue as demand grows. I might also transfer the cloud computing capabilities of our conventional data centers to Matsue.
At present, up to 24 containers can be installed at the Matsue site. One container can hold either nine or 10 racks (i.e., 360 or 400 servers, respectively), so the maximum server capacity is 9,600. Using this capacity to our advantage, we can reduce the operating cost and offer data center services at more reasonable fees.

So the location of servers can affect service fees too?

Suzuki：Yes, because electricity costs differ by locality, as do construction and land procurement costs. We are also planning to build a network of data centers so that, should an accident occur at one data center, others can cover for it without disrupting service. Plans to build similar data center parks elsewhere are already underway. We intend to construct several such data center parks in Japan, which will be energy efficient and ideal for a cloud computing environment. These will be connected to form a network and used as one virtual data center.

I imagine you’ve had untold difficulties in setting up the Matsue Data Center Park because it was the first of its kind in Japan.

Suzuki：Not really. It was challenging, of course, because we were the first to do it in Japan, but we were determined to solve any problems at hand, and we did. In particular, we set the goal of achieving world-class PUE levels because development engineers love intriguing challenges.

I hear a leading American container module manufacturer sent representatives to Matsue who observed the site and exclaimed that no commercial application like this existed anywhere else in the world.

Suzuki： I know some containerized data centers are owned by big corporations like Google for internal use. But ours is probably one of very few in the world that are used for commercial services. As we continue with our project, I hope to keep benefiting from the experience and expertise of CM Plus.