When Ed Milstein graduated from school as a chemical engineer, he landed a job for $15 an hour, writing a computer program to monitor the condition of bananas being shipped to New Jersey. Now, Ed wasn't a programmer—just for fun, he'd done some programming designed to gather data—but apparently his predecessor had cheated and written a program that generated random conditions, all within the tolerance required by the banana shippers. So Ed arrived on the scene just in time to land a great job.
Today, as founder and president of Network Intensive (and Compute Intensive),
Ed Milstein runs the multi-million dollar company that boasts more POPs in the basin than any other provider. He has a solid reputation among the business community for understanding the needs of business, from networking to connection to the World Wide Web. It looks like Ed's still got that great sense of timing—that ability to know when to take a risk—along with the know-how to back up his actions. This may be the story of his life, as he continues to learn and grow Network Intensive.
WWWiz: What's the difference between you and everyone else?
E.M.: There's a few ways to look at it. One is that we come from a different place. A lot of our competitors were born to be an Internet provider, and have certain constraints put upon them because of that. We came from the network consulting world, and still do a tremendous amount of that today. We started with a great amount of expertise in networking, and basically got into the business because we had been asked by a number of customers to hook them up to NETCOM, to do the local hook-up to the network, to do the whole security thing, and we found that the people we were dealing with at the other end were very inept, and we thought this was a place where we could provide our expertise.
The other component is, I was always felt there was this real niche out there for quality. There's always a high-quality provider, whatever industry you're in. If you're talking bout automobiles, it's Mercedes, Lexus, now there's Infiniti, everyone wants to go play in the high-value space. I think the same thing is true in the ISP business. I always tell people, "Think of us as Lexus."
We have about a 40% conversion rate. Look at our new customers every month. About 40% are coming from other ISPs. They cover the spectrum, all the way from AOL to local competitors. We even convert customers from MCI, places like that, but the only provider that we don't ever see conversions from is like UUNet or Cerf, people who we think are high quality also.
WWWiz: What's your mix? Are you mostly business and networks, or are you individual?
E.M.: We're the most expensive personal account in the basin, by a wide margin. The industry seems to have settled on a "twenty-dollar, all-you-can-eat" model. I think that's a bad model. I always use the analogy of the Blue Angels. A number of years ago, the Blue Angels had a major crash, and it was because the leader messed up, and the guys were so trained to follow him, they followed him to the ground. I always tell people we need to not follow the people who are going to crash into the ground. I think twenty dollars unlimited is a price where you can't make money, number one. Or if you can make money, it means low service, low quality. It means you run a higher ratio of modems, and a lower ratio of tech support, a lower ratio of bandwidth, and those types of things. So on personals, we do very well. We're $35 a month for a non-unlimited account, and last month's sales were almost 35% over goal in the personal area, and our goals are high.
WWWiz: How did you decide to start the business in general?
E.M.: I've always wanted to have my own business, probably from my college days. I'm actually a chemical engineer by education, not a computer scientist. To make a very long story short, I ended up stumbling into a $15-an-hour contract when I was a sophomore in college, doing programming. Fifteen bucks an hour, as a sophomore in college, could've been a million dollars an hour. That's the way it felt to me. And I ended up doing consulting for most of my undergraduate. When I graduated, I then had so much experience in computers, that I actually was in more demand. I wound up taking a job at Hughes Aircraft.
In about the early eighties, I started formulating my plan. I went to work for Sun Microsystem; I literally thought I would be there about a year, I would learn about the business side, the sales side, network with all the people doing high technology. I figured in about a year I could start my business; I met a lot of people who are here [at NI] today, and always told them I wanted to go start this business someday. Of course, no one ever believed me, because there are a million people out there who are going to someday start a business, right?
I ended up staying at Sun for six years. My salary eventually increased about 300%. They were very, very good to me; when my son was born I got a two-month, all-expenses-paid time off, because I actually wanted to quit, spend time with my newborn son. They said, "No, no, don't quit. Take two months off." Finally in about 1993, I said, "The time is now; I see a lot of opportunity out there."
WWWiz: So they almost prevented you from starting your business...
E.M.: No, they delayed it. My wife was actually my number-one advocate. When I was at about year three at Sun, she was telling me, "Quit. Go start your business." I said, no, I'm not ready. I want to go learn about this, and I want to go learn about this." And my original focus in 1993 was to be a software company, and to focus on multi-threading applications, which was my last job at Sun. The original name for the company, Compute Intensive, came from that original focus of "why would you multi-thread an application?" Because it was compute intensive. You don't go multi-thread IO-intensive applications or anything else, or graphics intensive, you go and you multi-thread compute-intensive applications. We did some work for Sub Pro, we did the development arm of Sun, we did some work for a large telephony company in Orange, and the multi-threading business helped get the company rolling. At the time, however, I saw that there were other opportunities, opportunities that were bigger, so I changed direction over time.
WWWiz: Where did you go to school?
E.M.: New Jersey Institute of Technology.
WWWiz: So you're from New Jersey?
E.M.: Actually I grew up in New York. I went to Jersey for a lot of good reasons, and JIT was a good school to go to, because it was only 6,000 people, and it was engineering only. So I think I got a good education there. The school was not highly theoretical. To give you an example, one of my toughest classes was a materials handling-type class, and the final was only five questions, and three of them had no right answer, which the teacher didn't tell us about ahead of time. Really, what he was looking for was "Okay, guys—you're out there in the real world, you're going to meet intractable problems out there, what do you do?" And it was that training that helped prepare me to be an entrepreneur, much more than "I'm an EE, and I've got to do this algorithm to come up with a circuit." That's good work, high-value work, and intellectual work, but it wasn't problem-solving skills. And so I was actually very thankful for my undergraduate education. I'm a little different. Not too many chemical engineers are running an ISP out there, or running a computer network integration company or anything like that.
WWWiz: Back to the battle of the flat rates, what do you think is the future of ISPs?
E.M.: I think the business is going to segment. I think there's going to be a portion of the business that focuses on the personal account, that is, the unlimited model. That market, I think most people today recognize, is going to commoditize, and that has never been a market that I've been interested in.
WWWiz: Do you have any plans to provide software on the Web?
E.M.: There's two ways to answer that question. One is that that's not a business that I'm terribly interested in, for a few reasons. One is, I think the competitors are too formidable, and too entrenched. I don't mind competing with MCI and Sprint on the ISP side, because they're figuring it out, just like I am. Okay? I don't think Microsoft's figuring anything out. I think they pretty much get it. [laughs] So, that the competitors are really too formidable, and will predominate in the space is one reason. Number two is, that's not a short-term way to make money. I have the belief that technology moves very slowly. You can have the most compelling technology in the world, it can be faster, cheaper, better, nicer, easier, and people still don't adopt it overnight. It still takes years, because the technology world has a great deal of inertia behind it. I can give you lots of examples: MS-DOS still sells. You go down to Egghead, they have a big inventory of MS-DOS, and people buy it every day. That's a perfect example. Nobody should be buying DOS any more! Nobody! But they still sell a ton of it! They thought Windows 95 was going to take over the world. They've sold 20 million copies, but that's only a percentage of the world. And this was the biggest thing to hit the world since they figured out how to slice bread.
I just don't see it as a short term. One of the things that I think this company is good at, going back to Compute Intensive and where we started, is I think we're small enough, and we're bright enough, and that we act quickly, if that really does start to become an interesting market, we'll react to it. Is it something that I really want to bet the company on in '96, '97, or maybe '98? No, probably not. Are people going to go and be successful there? Absolutely. I think when we start to see some inroads, meaning when we even start to see tools for the Web, that are built like that, that is a good number of years away before we start to see tools for our word processor that work that way.
So now, when I want to go type you a letter and say, "Thanks for coming by; I really enjoyed the interview," I've gotta wait to download something from the slow Microsoft site? I don't think so. I think, yeah. Someday. A year from now, two years from now? No. Should Microsoft be looking that way? Absolutely, because if not, there'll be a new Netscape that does. And if one of those guys doesn't figure it out, there'll be a third Netscape that does.
WWWiz: So where do you think this is all going?
E.M.: I think most of my viewpoints on the future are fairly well understood. I do see the splitting, the segmenting of the industry, where the personals commoditize, and there's the second view where what people expect from an ISP or network integration company is going to go way up. I think commerce is clearly going to be the biggest thing on the Internet. I think that will really predominate for the next year or two.
A "backbone" is the main network which is used to interconnect other networks. In the case of the Internet there was, until recently, a backbone which was funded by the National Science Foundation (NSF). The NSF backbone was used by university, government and regional networks to communicate with each other. The NSF backbone provided these networks with a data communications service that crossed the country. The backbone users only had to connect to the backbone wherever it came closest to them geographically. The NSF provided network would then transport their data to the remote network, whether nearby or across the country.
Acceptable Use Policies
Because the backbone was funded by the NSF, only certain types of data were allowed. Usually the data was restricted to material that was of an educational or research-related nature. Commercial traffic was prohibited since the NSF did not wish to use taxpayer money to provide network services for commercial organizations.
This "acceptable use policy" imposed by the NSF led organizations that wanted to do commercially oriented things on the Internet (like marketing or data transfer between businesses) to form an organization known as the Commercial Internet Exchange (CIX). CIX created an "exchange point" where providers of networking services could connect to each other without using the NSF backbone. Traffic which traveled between two organizations using the CIX was not subject to any usage policies imposed by the NSF. Thus, non-research traffic was acceptable, and the Internet started to become a place where businesses could do commerce.
The Internet Goes Commercial
At the end of April 1995, NSF funding for Internet backbone services ended. After that date, the many network service providers, universities and government agencies that used to rely on the NSF backbone started using a new architecture. With this new architecture, the trend begun by CIX was carried to its conclusion and restrictions on commercial traffic essentially came to an end.
Instead of having a single organization provide backbone services, the NSF created an interconnection strategy which consisted of multiple exchange points, like CIX, where Internet service providers could exchange traffic with each other. Initially, a small number of NSF "Network Access Points" (NAPs) were defined. Eventually the NAPs proved valuable enough that additional NAPs were created. The NAPs are distributed on a geographic basis, and the original four were in Pennsauken, NJ, Chicago, near Washington, DC, and the San Francisco Bay Area.
At each exchange point, a high-speed local area network is used to interconnect the networks of the participating ISPs. If service provider A has a customer who wishes to send data to a customer of service provider B, then ISP A will use his own network to get his customer's data to the exchange point. Then he will "hand off" the data to service provider B. From there, ISP B uses his own network to deliver the data to his customer.
Any traffic which must go across the country (or around the world) is carried by the ISPs themselves, on their private backbones. Because of this, there are really multiple Internet backbones belonging to each of the major network service providers. Regional and smaller providers will often pay the larger ISPs for the right to use their backbones. Since one of the goals of the Internet is to make sure that all the users connected to it can reach each other, the exchange points provide a critical service by allowing the customers of one ISP to communicate with the customers of other ISPs.
The reality of the Internet today is that there is no single backbone. Instead there are multiple exchange points.