I/O: Thoughts from Digital World
‘Highways’ may be the wrong metaphor
Jonathan Seybold is publisher of Digital Media and CEO of Seybold Seminars, which produces the annual Digital World conference.>>
The metaphor that everyone has been using for the communication structure of the future is “information highways.” This may have been appropriate at last year’s Digital World, when we were talking about relatively “dumb” digital cable TV systems with lots of channels that really did move programming down a wire much the same way that a car moves along the highway.
But it hit me somewhere between Bob Carberry from IBM demonstrating the applications of a gigabit network, Geoff Holmes of Time Warner explaining the incredibly ambitious prototype system his company is installing in Orlando, Larry Ellison talking about server architecture (in between his sales pitches for Oracle), and Dick Green of CableLabs describing the full-service network the cable industry is now planning to put in place. What we’re all really talking about is creating a single, giant, networked computer system that’s 100 to 1,000 times more powerful than any installed in the most advanced businesses — and (eventually) connecting the entire population to that network!
The magnitude of this task is worth pondering for a moment. There we were at Digital World, blithely discussing a world where TV settop boxes have the raw computing power of today’s top engineering workstations, and which are transparently connected (via fully switched networks capable of data transfer speeds 100 or more times greater than those of the best office systems) to massive digital file servers the likes of which we have never seen.
And we are talking about delivering this technology to every consumer that wants it and can pay for it.
When you start thinking about this network as a mass of interconnected, interactive computer systems instead of only thinking about it as a data highway, several observations spring immediately to mind.
1. It’s a daunting job.>> In the first place, thinking of these “highways” as computer systems brings home the magnitude and the difficulty of the undertaking. These systems will have a completely different character from any computer systems we have built before. They must combine unprecedented scale (in terms of numbers of users and the amount of data being handled), extraordinary robustness, exceptional security, and unprecedented ease of use. This is a pretty daunting task!
It is ridiculous to assume that the new generation of interactive services will be any different. It takes a while to get to “critical mass” (both you and the person you want to communicate with have to have access to the service), but once this happens, it is logical to expect that we will see widespread use of real-time video telephony, video “voice mail,” electronic “package delivery” of multimedia information, interactive forums and “gathering places,” and all types of virtual games in which people interact in cyberspace.
People who have been exposed to the existing, often primitive computer interactive services talk (and worry) about this kind of application all the time. But the people who are actually building the new systems seem largely oblivious to it.
2. It looks a lot like the phone system.>> With the crucial exception of the amount of data we will be moving around, the sort of super high-speed switched system everyone is now talking about bears a much stronger resemblance to the existing phone system than to the existing cable TV system.
This, of course, is why almost everyone assumed initially that it had to be built up from the phone system. It was only a couple of years ago that people began to notice that the really expensive part of the system is the physical wire — and cable already has high-bandwidth wiring in place.
What really complicates things is that cable and the phone companies currently operate under such vastly different regulatory and economic conditions. The cable people do not have to worry about things such as universal service and prohibitions on deriving revenue from the content they carry. The phone companies do.
3. The computer industry is an essential player.>> If you are going to build massive interactive computer systems, you are almost certainly going to need to involve the people who build computers and computer systems, especially those who understand the special needs of networked systems.
If you turn this around and look at it from the standpoint of the computer industry, it is instantly clear that this is the big opportunity of the coming decades. The market is likely to dwarf anything the industry has seen to date. No wonder Apple, IBM, Microsoft, Oracle, Silicon Graphics, et al. are suddenly paying so much attention to it. But the rest of the computer industry has been remarkably slow to catch on.
4. The fundamental importance of open systems.>> If the computer industry has one lesson to teach the cable, telco and consumer electronics industries, it is the importance of open systems — the importance of an open, competitive market where a wide variety of companies, large and small, can compete to provide the best solution for some component of the overall system.
The experience of the computer industry should also teach us that however much one may profess to believe in open systems, no supplier will willingly give up competitive advantage. Open systems have come about because users insisted on them, not because the vendors altruistically decided they were a Good Thing.
If we think about the new systems as giant interactive computer systems it is pretty clear what we need to ensure competition:
• Standard network protocols and data formats that make it easy to connect one network to another and/or to connect all manner of devices to the network. As long as protocols and data formats are described and published, any network should be able to link to (or compete with) any other. In the end, users should not have to care whose wires (or radio waves) are carrying their data.
• A standard network interface for settop boxes. There is no need to standardize the hardware or software within the box. There can be plenty of room for competition — and innovation — here. But it is important that the interface be completely open so that anyone can compete. It is also important that there be a small number of network interfaces (preferably no more than two), so that we can create a huge nationwide (or even worldwide) market for the consumer devices that connect to the network.
• A standard network interface and command protocols for information servers. We want a system that will allow anyone to connect into the network with any size or type of server. As long as you build a server that has the proper data interfaces and understands the command protocols it must deal with, it should work on the network.
• Standard transaction protocols. In order for the system to be truly open, anyone who wishes to must be able to conduct business over it.
What is not yet clear is how we get these standards. My hope is that the network carriers (the cable, telco and wireless people) will understand that it is in their interest (as well as ours) to create open standards. They are in a position to insist on this in their dealings with the computer industry people and with each other. However, human nature (and the profit motive) being what it is, the network carrier people are going to be looking for ways to ensure that their position in all of this is as lucrative (and protected) as possible.
5. Likely products and services.>> Movies on demand, interactive catalog shopping and interactive classified ads are easy targets because they all represent very large markets that could (presumably) be better served by online interactive services than by current approaches and technology.
But if we think of what we are building as giant interactive computer systems, other obvious applications jump immediately to mind: Probably the most pervasive use of the existing interactive computer systems is person-to-person communication. The examples are all around us: electronic mail, electronic bulletin boards and information services, Minitel services in France, and, of course, the use of the current phone system.
It is ridiculous to assume that the new generation of interactive services will be any different. It takes a while to get to “critical mass” (both you and the person you want to communicate with have to have access to the service), but once this happens, it is logical to expect that we will see widespread use of real-time video telephony, video “voice mail,” electronic “package delivery” of multimedia information, interactive forums and “gathering places,” and all types of virtual games in which people interact in cyberspace.
People who have been exposed to the existing, often primitive computer interactive services talk (and worry) about this kind of application all the time. But the people who are actually building the new systems seem largely oblivious to it.
INFORMATION HIGHWAYS VS. COMPUTER SYSTEMS
The “information highway” is probably still a useful metaphor. It reminds us that we are setting out to build an infrastructure that will be used to “transport” all manner of goods and services — one which is almost certain to transform the way we are informed and entertained, and the way we do business.
However, it is vitally important to remind ourselves that what we are really going to be building are really not so much “highways” as interactive computer systems of a truly awesome magnitude. We are embarking on a massive effort to put an ever-increasing proportion of our population “online.” This is going to take decades. It is going to involve wrestling with profound social, commercial, technical and human issues. But it looks as if it really is going to happen. As far as I am concerned, the best news out of this year’s Digital World conference was the clear overriding concern that we pay attention to what we are doing and do our best to do it right.
Jonathan Seybold