IPv6 - MP3s & presentations This site has PPTs and MP3s from last year's IPv6 Conference. IPv6 is the next generation of Internet IP addressing. It is rapidly becoming necessary to replace the old IPv4 numbering scheme because we are running out of unique numbers with the explosion of devices such as cell phones and smart appliances. Although many of the presentations are technical, be sure to check out Latif Ladid's opening presentation. It is full of gorgeous mockups of future smart devices.

Byte > Column > P2P Meme > Presence Management, Instant Intranets > September 22, 2000 This Byte artice is a really good, basic, overview of some of the technologies that are being lumped into peer-to-peer (P2P). It's pretty even handed and not sensational. I'm particularly struck by the recollection that, once upon a long time ago, the Internet was entirely a peer network. Then these firewalls and NATs (Network Address Translation) and dynamic IP address assignment schemes got in the way. Napster returned us to the peer Internet, allowing any user to directly (well, almost directly. There's still the matter of routers and firewalls -- but no DNS problems) contact another and share resources. I've finished my Hive Computing white paper and am starting on a white paper on the other aspects of the P2P phenomenon. It's pretty fascinating stuff.

DataSynapse, Inc. I'm struck as I read about DataSynapse's architecture that this is not new at all. That's true of the whole P2P thingie; these concepts have been around for a long time. But the network diagram at the DataSynapse site pretty much describes what we built at ACNielsen starting in 1989. We called it the Automated LAN Production System (ALPS), and we set it up in the Fond du Lac, WI production site in spring 1990. Using DOS technology and the Clipper development language, our consultant, Geneer, created a peer-to-peer system in which multiple worker PCs polled a central database for work to do. Each PC was characterized by its "capas", short for capabilities. Some could run Harvard Graphics in batch, others could scrape data down from the mainframe into Nielsen's data analysis tool at the time, SCAN*FACT PC. Others could create Lotus 1-2-3 files or proprietary MegaChart charts from the data. Still others could print the results in nice, page-numbered, books. A central console allowed the operator to stop, start, or reconfigure the worker PCs. This was P2P, and it resembled DataSynapse's scheme very closely. We knew what we had done was pretty special, but, as so often happens in a large company, few top managers knew the value of it. Using this system, a staff of 17 could produce 5 million customer pages of output a year with a dozen 486 PCs or so. We wanted to take the concept further and make use of idle cycles across the enterprise, but top managers were satisfied with the substantial savings and productivity we produced with the existing system. If we only knew we were just ahead of our time!