Blunder Dome Sighting  

Hangout for experimental confirmation and demonstration of software, computing, and networking. The exercises don't always work out. The professor is a bumbler and the laboratory assistant is a skanky dufus.

Click for Blog Feed
Blog Feed

Recent Items
To Engineer is to Tinker?
A Feed Too Far
Security is a Programming Problem?
Ending the Madness: Deja Triple Vu
Just Ducky, Simply Ducky. And You?
All Clear: End of Test #1
Why Learn Assembly Language
Linking the Atom Feeds
Button, Button, Where's the Update?

This page is powered by Blogger. Isn't yours?

Locations of visitors to this site
visits to Orcmid's Lair pages

The nfoCentrale Blog Conclave
Millennia Antica: The Kiln Sitter's Diary
nfoWorks: Pursuing Harmony
Numbering Peano
Orcmid's Lair
Orcmid's Live Hideout
Prof. von Clueless in the Blunder Dome
Spanner Wingnut's Muddleware Lab (experimental)

nfoCentrale Associated Sites
DMA: The Document Management Alliance
DMware: Document Management Interoperability Exchange
Millennia Antica Pottery
The Miser Project
nfoCentrale: the Anchor Site
nfoWare: Information Processing Technology
nfoWorks: Tools for Document Interoperability
NuovoDoc: Design for Document System Interoperability
ODMA Interoperability Exchange
Orcmid's Lair
TROST: Open-System Trustworthiness



Lost in Twisty Overlays All the Same: Peer Pressure

I have been clipping on a lot of topics, and now want to organize them where I can see the result and also keep down the number of individual posts.  This may be contrary to blog etiquette, although I find it disheartening to trace one-line feed articles to one-line blog entries that are nothing but another link.  I don't mind that with Scoble's link blog, because I know what that is.  Whatever.  I am consolidating clippings into posts as a way to give me an organization that I can handle.  More or less ...

Peering Network for Content Updates: FeedMesh

Sam Ruby: FeedMesh.  There is an immediate and practical problem in the blogosphere around the collapse of syndication.  As was discovered on MSDN, certain approaches to syndication do not scale, being vulnerable to a form of commons-saturation crumble.  The hosts become effectively DoS-ed by the incessant demand for pulls of the latest feed.  This is partly attributable to a mistake in the MSDN approach, but it also comes from the fact that syndication is a pull technology (its great advantage, in fact) and the compulsive appetite of feed readers for polling the site constantly for updates to the feed. At this weekend's Friends-of-O'Reilly (Foo) camp, illuminati of the feed-syndication community have been putting their heads together about this problem and looking for a scalable remedy.  It looks pretty bottom-up at the moment, and it is awesome what happens when some magnet draws these folks into one place. This article provides valuable links to the connected parties and to places where the initiative is being discussed.  And you get to see Sam Ruby tell Dare Obasanjo to stop trolling.  The FeedMesh, uh topic, is posted on the Foo camp Kwiki, so it might take off in all unruly directions at once.

Efficiently Finding Resources the Peer-to-Peer Way

ACM News Service: Cooperative Search Technology.  One important aspect of peer-to-peer networks is the ability to search for resources.  This can be of great importance to grid computations and overlays where computing resources of various kinds are shared.  The discovery and location process, and ways for it to work among small specialized groupings while scaling with the overall population of sites is critical. The UCLA algorithm (I wonder when it will get a better name than that) appears to be extremely network-friendly while offering these advantages.  A software library is being built for incorporation of the search algorithm into applications "in about one or two years."  The work is sponsored by NSF and DARPA and follows on earlier work at Stanford and Hewlett Packard in 2001. The Kimblerly Patch 2004-09-08 Technology Research News article provides a sketch of the method.  Although it seems to involve extensive contact with nodes, the method assures that the number of nodes touched in a search grows more slowly than the network itself, and models with 100-million nodes are considered.  The breakthrough is in having the process implemented locally, through messages passed among neighbors only.  That makes me curious about hwo the network is itself discovered, and for that one apparently needs to look up BruNet, a P2P network that operates by local and simple actions.  That's exactly what I am looking for in distributing Miser and other distributed-object systems. The key technical reference is apparently "Scalable Percolation Search in Power Law Networks" on Arxiv.

Semi-Centralized P2P

ACM News Service: Building Peer-to-Peer Applications.  This is another EU project to develop open-source solutions for important technologies.  This P2P Architect project supports important commercial activities.  I am looking for fully decentralized operation, but the fact that this model supports full collaborative editing of the same objects has my attention. The 2004-08-20 IST Results feature article provides more information.  There is a link to the P2P Architect project and more detail on the participating organizations.  [The linked page was broken on my first visit, and it remains so now.  There seems to be great absence of the usual project materials that accompanies completed IST Results and I find that rather strange, especially with the avowed open-source approach. -- dh:2004-09-11]

NOMAD Middleware for Roaming

ACM News Service: New Middleware Platform for Roaming Mobile Users.  NOMAD is a project of the Information Societies Technologies (IST) program being conducted in the European Union.  There are many of these projects beginning to mature under IST, and NOMAD is one of them for mobile-user discovery and collaboration. The 2004-08-13 IST Results feature article provides more background and links to the project on Integrated Networks for Seamless and Transparent Service Discovery.  Although location-aware end-points are featured, I hold onto this link because this seems like yet-another view on discovery that may be as useful for fixed (or NAT-ed) endpoints as location-aware ones.

P2P Risk Exposures

ACM News Service: P2P Drag on Nets Getting Worse.  I find it amazing that the commercial sides of P2P businesses sell caching systems to ISPs so that the P2P glut on networks is relieved.  Beside creating significant traffic (30 to 70%), there is a significant risk to intranets by P2P traffic because of malware transmission and also proprietary materials that may expose organizations to lawsuits.  Finally, the ease with which P2P provides anonymity and secrecy creates an exposure to fraud with the cooperation of insiders. The Carolyn Duffy Marsan 2004-08-02 Network World Fusion article provides useful links along with discussion of the prevalent P2P technologies as well as countermeasures employed in businesses. A related interview (audio) and link page focuses on BitTorrent, a current favorite.

Sharing Lightens the Download

ACM News Service: Sharing Lightens the Download.  This article looks at the latest torrent/swarm-based P2P arrangements as powerful for legitimate multi-media and software distribution.  The BBC is experimenting with the mechanism for downloading TV broadcasts.  The Chord effort is found appealing for useful distributed storage and the prospect of preservation via distribution.  [The Kieren McCarthy 2004-06-26 New Scientist article is apparently not available on-line.] In researching this lead, I found that there are a variety of resources on this and related topics, including the proceedings of the International Peer To Peer Systems (IPTPS) workshops, IPTPS02, IPTPS03 and IPTPS04.

Construction Structure (Hard Hat Area) You are navigating Orcmid's Lair.

template created 2004-06-17-20:01 -0700 (pdt) by orcmid
$$Author: Orcmid $
$$Date: 10-04-30 22:33 $
$$Revision: 21 $