Some politicking around the P2P Working Group.
>>+ Peer-to-Peer:The Federation Fights God
>> By Lloyd Blythen
>>"The world is moving towards less-centrally, less-hierarchically
>>organized decision making, towards more-distributed, collaborative
>>decision making. Peer-to-peer reflects a socio-political change"
>>Yeah, right. Brian Morrow may be the only person we've met lately who
>>thinks leaders are about to relax their hold on the reins. More likely we
>>can expect some hyperbole from the newly-elected chairman of the
>>Peer-to-Peer Working Group - who also happens to be president and COO of
>>Endeavors Technology Inc, a P2P software vendor based in Irvine, California.
>>
>>If Morrow sounds a little desperate, it may be because the Working Group
>>hasn't made much progress to date and faces an uphill fight. "Since last
>>summer [the Group was formed in August] not much was accomplished within
>>the Peer-to-Peer Working Group. I want to turn that around quite
>>drastically," he told Unigram.X last week.
>>The Group has a problem. It was launched by P2P advocate Intel Corp and
>>now has a complement of some 200 industry members. But at its three
>>meetings to date, three highly influential industry players have
>>resolutely refused to join up: Morrow euphemistically describes them as
>>"attending but undeclared", and diplomatically sidesteps our speculation
>>that IBM, Sun and Microsoft are the companies in question.
>>
>>"It's a small-'p' political challenge for me that the major players must
>>be persuaded to join or risk being out of the mainstream," he says. The
>>major players may consider that an acceptable risk, however, considering
>>that they represent the bulk of the software market - Java, legacy,
>>Windows - and, Intel's weight aside, need not view the Group as much of a
>>threat.
>>
>>But there's a bigger issue which binds them together: God. "Who gets to
>>play God?" is the key question facing the P2P community, according to
>>Morrow, by which he means who will own the registry of peers. The most
>>likely answer at present may be "America OnLine".
>>The beauty of P2P is that, unlike the alternative client/server
>>architecture, all peers have the ability to perform both client and
>>server functions. The down-side is that, on first connecting to the
>>network, each peer must blindly search for other peers - unless something
>>"plays God". In this context, God is a central registry to which peers
>>announce themselves and obtain information about other peers that are
>>on-line; it may also act as a repository of security and other
>>information. At present AOL has by far the most pervasive P2P network,
>>and is therefore in the best position to play God, thanks to its
>>ubiquitous Instant Messenger service.
>>
>>The Peer-to-Peer Working Group thinks it has a better way, however. Using
>>existing standards rather than developing its own, it aims to promote a
>>"federation" architecture to rival AOL's hierarchical structure. Within a
>>federation, typically up to 1,500 local peers share a "demi-god"; many of
>>these can be interconnected to make their data more widely available and
>>to extend the reach of the overall P2P network.
>>
>>Morrow recognizes that his Group needs to promote its federation model -
>>and, with his Endeavors hat on, that his company's Magi event-driven
>>platform needs to make some major sales - in a hurry if it is to stave
>>off AOL's own P2P efforts. The Working Group is focusing on niche
>>technologies - such as firewalls - where existing standards can be
>>employed or extended to support P2P. He has set a six-month target, not
>>for achieving a certain number of sales or seats but for "getting the
>>message across and getting a few big sponsors."
>>
>>Just who those sponsors should be will depend on what applications are
>>both popular and amenable to P2P. AOL's AIM network is used heavily for
>>on-line collaboration, which - perhaps unfortunately for the Group - is
>>an area that needs to be focused on.
>>In addition to file-sharing and collaboration, however, Morrow identifies
>>two other promising areas in which to promote P2P: gaming and real-time
>>streaming; and CPU cycle-sharing initiatives such as the Grid projects.
>>He also describes an e-business model that maps onto federated P2P very
>>closely: secondary e-commerce, where a large vendor initially supplies
>>several smaller ones, which in turn trade among themselves as peers to
>>avoid individually holding large inventories.
>>
>>It remains to be seen whether the Peer-to-Peer Working Group can achieve
>>critical mass with the federated model, but Endeavors is determined to
>>contribute. Its existing Java software will be ported to the open-source
>>Python language, and the company has announced an open-source Python
>>Virtual Machine to run it (see story in this edition). The VM has been
>>trimmed by an impressive factor of about eight, from a core obtained when
>>Endeavors hired several researchers from the University of California at
>>Irvine: it now occupies less than 200Kb and requires just 64Kb of RAM for
>>its heap. Endeavors aims to "start with the Palm and go down," says
>>Morrow, and one assumes he's not waxing prophetic. He predicts an
>>eventual mass market for P2P devices as small as air-conditioning
>>thermostats, but first, he says, "we want to be inside the cellphone."
>>
-- Gregory Alan Bolcer | gbolcer@endtech.com | work: 949.833.2800 Chief Technology Officer | http://www.endtech.com | cell: 714.928.5476 Endeavors Technology, Inc. | efax: 603.994.0516 | wap: 949.278.2805
This archive was generated by hypermail 2b29 : Fri Apr 27 2001 - 23:18:25 PDT