[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Gzz] 11th (hh)
From: |
Tuomas Lukka |
Subject: |
Re: [Gzz] 11th (hh) |
Date: |
Sat, 16 Nov 2002 18:47:14 +0200 |
User-agent: |
Mutt/1.4i |
On Mon, Nov 11, 2002 at 08:22:24PM +0100, Benja Fallenstein wrote:
>
> Hi Hermanni,
>
> address@hidden wrote:
>
> >-DSHT (distributed sloppy hash table). Anyone aware of this ? This method
> >should
> >solve hotspot problem related to DHTs (could solve other issues also).
> >
>
> No, but I've read what you've forwarded. Sounds good, but my instinct is
> to wait until more people are working on it before thinking too much
> about it ;-)
Exactly. Hemppah: remember that when using non-peer-reviewed sources, there
are a great deal of "brilliant" inventions that, on closer inspection,
are not quite so brilliant.
Even in peer-reviewed stuff, there are "great inventions" that get no second
mention in the literature, because there's something wrong with them.
Be very careful about what you trust.
> Preliminarily: I think DSHTs are obviously great for getting data by its
> hash.
"Obviously": I'd not go even as far as that; I'd say they are something which
does need to be examined but whether they're great will be seen later.
> I'm not sure yet about distributed indexing, which is also
> important; for example, in Gzz, we need a "find all links to and
> transclusions of this piece of data" query. Here, it does not suffice to
> find a single link/transclusion; we need to find all. DSHTs will
> probably not help here.
Yes. We have some quite interesting problems...
Tuomas