RE: [P2PSIP] Term for the peer who is responsible for a specific key

   Abundant info!  No more supplement.
   In fact, I do not care "root peer" or "responsible peer" for the term about the peer which has the responsibility to
store a specific key; I think that we need uniform and explicit definitions for those terms in one distinct glossary I-D or
the concept draft, which are useful and convenient to the contributors of P2PSIP WG.

--Hewen

-----Original Message-----
From: Henry Sinnreich [mailto:hsinnrei@adobe.com]
Sent: 2007?8?31? 0:37
To: Peter Pan; Philip Matthews; Zheng Hewen
Cc: Jan Seedorf; P2PSIP Mailing List
Subject: RE: [P2PSIP] Term for the peer who is responsible for a specific key

As Jan Seedorf has already indicated (see below), the root in a DHT is well defined and documented. Here are three examples,
including one from the P2P SIP WG I-D :-)>

If anything, this trail shows the need for a DHT and P2PSIP glossary.

Henry

A. Chord vn. 0.6.3 released on May 3, 2007 features:
* The root node of routing creates and removes replicas on a DHT instead of the requester if replication is enabled.
* In case of DHT put, get and remove, a node requests spare root candidates to be prepared for failure of the root node.

See http://overlayweaver.sourceforge.net/news/

B. Chord and Bamboo

"The DHT assigns every key in the identifier space to a node, which is called the root (or the successor) of the key. The
main primitive that DHTs support is lookup, in which a node can efficiently discover a key's root. The lookup protocol
greedily traverses the nodes of the DHT, progressing closer to the root of the key at each step."

See http://srhea.net/papers/ntr-worlds05.pdf

C. In the IETF P2P SIP WG:

"Req 7-5: The peer protocol MUST accommodate a DHT for a fast routing  algorithm that minimizes hop count to the root
node..."

See
http://tools.ietf.org/wg/p2psip/draft-bryan-p2psip-requirements-00.txt

Henry

-----Original Message-----
From: Peter Pan [mailto:huang-ming.pan@comcast.net]
Sent: Thursday, August 30, 2007 11:00 AM
To: Philip Matthews; Zheng Hewen
Cc: 'Jan Seedorf'; 'P2PSIP Mailing List'
Subject: Re: [P2PSIP] Term for the peer who is responsible for a specific key

neither have i, though i did heard of "root beer".

peter

----- Original Message -----
From: "Philip Matthews" <philip_matthews@magma.ca>
To: "Zheng Hewen" <hwzheng@huawei.com>
Cc: "'Jan Seedorf'" <Jan.Seedorf@netlab.nec.de>; "'P2PSIP Mailing List'"
<p2psip@ietf.org>
Sent: Thursday, August 30, 2007 8:45 AM
Subject: Re: [P2PSIP] Term for the peer who is responsible for a specific key

> I personally have never heard of the term "root peer" before, and I
> would never guess that it mean the peer that stores the content
> associated with a key. Can someone explain where this term comes from?
> What is it the root of?
>
> - Philip
>
> On 29-Aug-07, at 21:24 , Zheng Hewen wrote:
>
> > Hi,
> >
> > I suggest that the concept draft would adopt those terms - "P2PSIP
> > root peer" and "P2PSIP replica root peer", they are clear and useful
> > for contributors to P2PSIP WG.
> >
> > --Hewen
> >
> > -----Original Message-----
> > From: Jan Seedorf [mailto:Jan.Seedorf@netlab.nec.de]
> > Sent: 2007?8?29? 23:53
> > To: P2PSIP Mailing List
> > Subject: RE: [P2PSIP] Term for the peer who is responsible for a
> > specific key
> >
> > Dear Henry and Spencer (I am replying to the last two mails on this
> > thread),
> >
> > In the DHT literature the node responsible for storing the content
> > belonging to some key is usually referred to as the "root" node for
> > that particular key. If content for a key is additionally stored at
> > several nodes in the DHT to provide resilience on the DHT layer,
> > these nodes are referred to as the "replica roots" for that key.
> >
> > Hope this helps,
> >
> > - Jan Seedorf
> >

_______________________________________________
P2PSIP mailing list
P2PSIP@ietf.org
https://www1.ietf.org/mailman/listinfo/p2psip