A Practical Navigator for the Internet Economy

Internet Unsatisfactory For "Business Quality" E-Mail Says OMB Task Force --- Federal E-Mail Program Office Offers Electronic Gov't Based On OSI Standards & Defense Message System --- Move May Open Gov't Systems To Clipper-Like Encryption High Transition Costs Could Mean End Of --- Government Use Of Internet To Disseminate Information To Citizens

All great programs have modest beginnings. In August 1994 at OMB's behest, a federal-government, E-Mail Program Office was established. The E-Mail PMO is to develop a national system for moving "business quality" electronic information among all employees at all levels of government including state, local, and tribal. "Business quality" turns out to mean double key escrow encryption. (Tessera-Fortezza and MISSI are the new watch words. Dial 1-800 Go-MISSI at NSA to find out more.)

The goals are ambitious. The new system, in some quarters, is referred to as a means of delivery for electronic government. But what has this to do with the Internet? Well, it seems that OMB has decided that the Internet and the applications written for the upper layers of the TCP/IP protocol suite are not an acceptable vehicle for civilian government use because it is not a "business quality" operation. Furthermore an Electronic Mail Task Force has recommended, on behalf of OMB, that federal system be based upon the OSI suite of protocols and on the world wide, DoD, Defense Message System.

The Feds are behaving as though the stunning growth of the Internet and world wide dominance of TCP/IP never occurred. In a time of unprecedented budget cuts they are telling agencies to spend whatever it takes to build a new system of communication that is compatible with the existing Internet only at the transport and e- mail level. Web and Gopher? Forget them. There are no such tools in the OSI arsenal.

What's the price tag on the administration's folly? Perhaps 500 million dollars over the next five years. In May 95 Loral was awarded a 500 million dollar contract to begin the actual build out of DMS. Press accounts say that, if Loral sells the system to the rest of the government, its take could rise to over a billion. If the project goes further, one may expect that the only way agencies will be able to bear the costs is to cease their incompatible TCP/IP applications like Gopher and Web based dissemination of information to the public.

The push is on. On Friday July 28 a nine page document appeared on the DISA web server (http://www.itsi.disa.mil/dmshome.html). There DMS notes that it is rapidly advancing toward the delivery of business quality messaging needed by the rest of the government and that this, plus the endorsement of DMS specifications, by the "OMB-chartered E-Mail Task Force" make DMS a "serious candidate" for providing not only service between agencies but also "service to and between individual citizens or trading partners."

On March 9, 1995 DoD ordered its current system, AUTODIN, phased out as soon as possible. In this context the July 28 sales pitch says: "All federal departments and agencies currently using AUTODIN will need to address how they will continue to provide or receive similar service when AUTODIN is gone. This is an incentive to seriously consider transitioning their existing AUTODIN capability to the DMS." DMS conveniently forgets that it was cited in the May 8 Network World as saying the phase-out may take ten years. DMS hammers home a conclusion that government agencies have one of three choices (1) adopt DMS; (2) fund a DMS equivalent, or (3) seek "proprietary solutions for their business quality messaging." Notably use of the Internet is not offered as a fourth alternative! So too bad UUNET, PSI, NETCOM, Microsoft Network, Netscape, First Virtual, Commercenet. You have just been handed a vote of no confidence by the administration that not so long ago sang your heavenly praises. But have faith. There's always a new pot of gold. We hear that a sale of two million Fortezza cards is in the air!

Some unanswered questions: why can "business quality" e-mail be found only within the OSI suite and DoD specifications? Do we have a back door way of using this program to get key escrow encryption installed throughout civilian government? We intend to return to these issues.

Peering & Transit At The Naps And The Club Of Six

We return to the NAPs and look at how the issues of peering and transit are evolving there. We summarize the presentation on the significance of the Club or Six (ANS, Sprint, MCI, PSI, UUNET, AGIS/NET99) that we gave at the Telestrategies Conference two weeks ago. Basically there are six major backbone providers which connect at most of the NAPs and MAEs and which have agreed to peer with each other and without charging transit fees. Afterall, all have large backbones, so no one should need to point traffic across the backbone of another to reach a third party.

The sensitivity of this issue however hinges on what happens when another large national provider comes to the six and wants to connect. Currently the preferred method is for one of the six to make a peering agreement and sell transit (meaning peering access) to the other five. This effectively makes the large network a customer of one of the six.

But what if the large network wanted to peer directly with all six? We have been told by people with policy making authority that the six would discuss the pros and cons of accepting the would be joiner into the club. Then the six would either become seven, or the applicant would be told that it did not qualify for cost free peering with any of the six. In such a case, buying peering and transit from one would be its only alternative short of suing the six for restraint of trade. Precisely because of the legal implications of what might be portrayed as collusion among the club members, key insiders have told us that these policies are never openly discussed. The group practice as it has been described to us on background is not to make unilateral cost free peering arrangements. (We conclude with a list of those interconnecting at MAE East and West, Ameritech, Sprint, and CIX members. PAC Bell didn't get us a list of their members.)

National Information Infrastructure: The Dark Side In Washington State

We publish the intro and Educational Privacy parts of our Washington state study. The text takes readers through the Educational Reform Movement's hunger for data and shared data bases. Welcome to the world of electronic student dossiers and the feds coaching local citizens on how to overcome the confidentiality barrier by encouraging data sharing between local agencies.

A New Name: The COOK Report on Internet

We have dropped the "-> NREN" from our name. We realized some time ago that there would never be an NREN. Unfortunately until we realized it was confusing to some readers we never found, until now, the time to do anything about it.