My Beloved IP Is Getting Trashed
by Troy Dixler

Published April 2002

Abstract:

We're turning the Internet and IP into a virtual sewer. Now before you start throwing stuff, let me tell you where I'm coming from: I'm an IP bigot, and a devoted, long-standing member of North American Network Operators Group (NANOG) and the IETF.

But over the last year, I've become frustrated watching the IP "gurus" introduce a horde of new standards (many still "drafts"), technologies and services, all of which appear to serve a noble purpose: To make IP suitable for any and all types of applications.

We've seen this movie before; it was called ATM. You'll recall, ATM was going to bring about the utopia of convergence—the integration of voice, video and data. It didn't. While it delivered the service guarantees needed for different traffic types, those capabilities came at a high cost, high level of complexity and high overhead (i.e. cell tax). I've seen better movies.

Now the ATM missionaries, having conceded that the world is becoming all-IP, have taken the connection vs. connectionless debate to the IETF. Dozens of working groups, with hundreds of vendors, are working on new standards and technologies.

The industry is spinning under a "draft-of-the-week" regime. But all these various Layer 2 and Layer 3 proposals, each designed to solve a specific point problem, fail one critical test: They don't make for simpler IP networks.

Instead, they are creating confusion (read: sewage) at the network edge. It's becoming next to impossible for carriers to filter out the winners and focus in on a single architectural approach.

Let me be clear: Many of the new protocols and technologies, such as MPLS and QOS, are useful for creating more predictable traffic flows across a shared IP infrastructure (i.e., the Internet).

But that's also the problem. Resource sharing always involves compromise, and it lacks true predictability and guarantees. This is not only true for logical circuits (e.g., MPLS label switched paths) but also for physical assets like routers, where schemes like virtual routing have been proposed. Because these schemes are based on a fundamentally flawed approach—sharing critical routing resources—they don't result in stability for carriers; instead, they increase complexity.

The real problem is that IP, a connectionless protocol, was never developed to be the universal protocol. ATM was developed to serve that purpose and failed.

Please note: By downloading this information, you acknowledge that the sponsor(s) of this information may contact you, providing that they give you the option of opting out of further communications from them concerning this information.  Also, by your downloading this information, you agree that the information is for your personal use only and that this information may not be retransmitted to others or reposted on another web site.  Please encourage colleagues to download their own copy after registering at http://www.webtorials.com/reg/.