Comments on Open Access
As we build Internet connectivity atop the existing telecommunications infrastructure we need to assure that there is a separation between connectivity and the services that are built upon the connectivity.11-Nov-1999

Please send comments and feedback to "Bob Frankston".

You should also look at my more recent thoughts in my Contents vs Connectivity Essay.

Open Access and IP

I see the goal of open access as creating a marketplace that reflects the "IP Infrastructure".

The power of the Internet has come from separating the transport from the applications. The communications layer simply transports packets without imposing a policy such as guaranteed delivery. The result of this architecture has been to replace the telecommunications industry and its services with a commoditized connectivity. This has driven the prices down and has allowed an explosion of innovation.

This innovation takes place outside the network, at the end points. Unlike the traditional telecommunications industry which required that all parties within the infrastructure agree to a standard, this new information infrastructure simply requires cooperation between two parties at the end points. And this is without the need to make a major capital investment in an special infrastructure for a single purpose.

The traditional telecommunications industry has provided the services itself and charge for their value. The IP Infrastructure allows anyone to create a service and charges only for the cost of transporting the packets.

The Caller-ID (CLID) service is a great example. Two ATT cell phones can only exchange CLID information if every intermediary supports the CLID protocols. In the end-to-end world of the IP infrastructure, the phones are directly connected and can exchange CLID and any other information they choose to. It's also worth noting that Bell-Atlantic can charge $6/month to stop blocking (holding hostage) the CLID information for my wired phone. They are acting like a troll under the bridge charging me for passing over the common roadway!

Open Access is simply the separation of the transport from the services. This is a very simple idea. In fact, those of us used to "free speech" should be offended by the carrier's insistence on spying on our packets to tell us what use is legitimate and which is not. It's like saying that there is one phone line for talking and another one for singing Happy Birthday.

In fact, this is the way the marketplace will form over time as the providers discover that the business models for a transport company and a content company are very different. But this may take a long time and there will be a tendency to try to lock customers into the provider's services in order to justify the large investments made.

The split between transport and service should be at a network level rather than the physical level. We see the consequences of these problems in the world of DSL. Since DSL technologies were created in order to run interactive TV over existing phone wires, there must be old style copper wire from the subscriber all the way back to the central office in order to provide the service because that was the design point. Modern and more economical facilities prevent one from getting DSL. For example, if the phone company has used fiber to your neighborhood then they cannot give you DSL! Since the telephone companies are in both the service and transport business they have an incentive to place major impediments in the way of competitors.

A Simple Proposal

We need to separate the concept of Internet Access from Internet Services. The word Access is not precise since we are talking about providing connectivity, which is simply the transport of IP packets to and from the subscriber's premises to an Internet backbone.

There are many services that can be provided including email, web hosting and "portal" services. These are all optional. One can run one's own email server. When you type a URL into your browser, you are going directly to the server without going through a service provider.

But for many users, the main value of the service provider is the handholding and "warm fuzzies". The service providers (ISPs) themselves want the users so they can resell their "eyeballs". This is why AOL is so keen on being a provider.

Since the service provider is in the "content" business, the choice of service provider must be distinct from the access provider. In fact, one can already pay $9.95/month to AOL for nondialup access such as when you use it only via your cable modem or from work. The only problems are that this is in addition to the fee that one pays to the IAP (Internet Access Provider such as RoadRunner) and I must still call the IAP directly when there is a problem. The fee to the RoadRunner does include their services which are then redundent. Of course, one has the option of having multiple ISPs for different and even overlapping services.


What about Cable TV and local telephone services? Shouldn't they be included?

These are examples of the traditional telecommunications services that are defined by the provider. The IP Infrastructure is the means for opening them up. The only question is when will the infrastructure be capable of providing the services. This is happening rapidly with radio over the Internet being a viable alternative to broadcast. This will happen to telephony and then television over the next decade or two. One goal of Open Access is to assure that the current providers do not impede the transition of their closed services to the IP Infrastructure. There should already be sufficient performance to allow local telephone calls over IP without the delays of compression.

Aren't all the restrictions necessary to protect the integrity of the network?

No! In fact, they aren't generally enforced because they are not needed and as the network grows they will be even less important. There is a concern that networks designed for delivering one-way traffic are not configured to support full capacity in both directions but the current restrictions are inappropriate. Even ATT is advertising the use of the network for two-way video connections! They also miss the point of the IP infrastructure by focusing on a limited view of current applications. For example, confusing the Internet with the Web and thinking of the Web as a broadcast medium like TV. The Web is just one of many applications on the Internet. Email, Chat, file transfer, telephony and document sharing are among the applications that predate the web.

The access providers are also confused about their role. Instead of addressing issues of connectivity, they ask what operating system the user is running and attempt to dictate the design of home networks and even try to proscribe them. When they do provide home networking capabilities, they deign to provide a very small number of IP addresses, such as three. In fact, as one connects devices such as printers and devices, many addresses will be necessary.

What about commercial use?

Who should decide on what constitutes commercial use? Cable modem service and DSL are consumer-quality services. The whole notion that the provider determines what are good uses and bad uses of the connectivity is offensive and a legacy of a telecommunications world in which the government and providers were deeply involved in all aspects. The only reason for the distinction is to allow the providers to charge according to the value of the service, a practice that we are directly challenging! One justification is that the commercial user is more demanding but they can just provide a higher level of support. In fact, the consumer is likely to do make more use of services such as video streaming and put a larger load on the network. There is also the issue of hiding taxes everywhere in communications. These must be made explicit!

While the goals may be noble, such as funding schools, pursuing them by these means is counter-productive and has the effect of denying business and the economies the benefits of the technology. For example, policies which keep the legacy phone wires in place rather than encouraging more efficient use of the capacity.

Shouldn't the providers be able to prevent reselling?

My personal opinion is that such restrictions are inappropriately invasive. The business question is whether there is a significant threat to their revenue base. So far, the increased use has offset the increased efficiency of the network. If customers need to share the cost of the connection and are willing to put up with the problems in doing so, then it represents additional business.

The restrictions also raise other question. If I have a student renting a room, must the student get a separate account? There are probably precedents in commercial law allowing a seller to tell me how I can use a product or service once I have bought it but such restrictions require exceptional justification. Until this is shown to be a major problem, it seems inappropriate to impose such restrictions.

Is this just about Cable Modems?

The current discussion is focusing on cable modems because this issue is "settled" for DSL. But it really applies to both and the access rules for DSL should be changed to allow innovation in how connectivity is delivered. The IP interface should be the network jack in the house (or premises), rather than the physical characteristics of the wire. But doing so requires the same separation of transport and service that I require of the cable modem suppliers.

What about newer technologies.

My assumption is that successful services will, in fact, follow this model with new companies having to focus on which business they are in -- transport or service. Some bundled plays, such as the so-called "Internet Phones" that provide limited services rather than Internet access are going to fail. Or, more likely, the services themselves will become an uneconomical backwater as the IP connectivity becomes the real service whether the client is a PC or a specialized device. In fact, wireless and wired are the same with the main advantage of low speed wireless is mobility and high speed may be more economical for general connectivity to all devices including nonmobile systems.

Beyond Cable TV

Both Cable TV and Telephony have a long history of special treatment with both being viewed as "natural monopolies". For Cable TV this has meant trading services and fees for access to right-of-way. One example is the community access channel such as NewTV in my town of Newton. Such access is much better done via streaming media servers. New England Cable News has a good example of such a streaming server. Doing so will greatly increase the value of such community access and set an example for effective use of the medium.

The streaming server is one example of focusing on the value of the new infrastructure rather than just selling faster modem access. While the restrictions on the services limit the marketing position to a Faster/Better/Cheaper strategy, the real opportunity is to create exciting and valuable new services. Instead of focusing on restrictions and defending cross-subsidies, we all benefit -- the users, the providers, and the economy, by looking past these skirmishes to the new possibilities.

With Open Access as the separation of the transport from the content business, we allow everyone to add value and innovate. It is foolish and short-sighted for the providers to focus on defending turf rather than creating an exciting marketplace.