sks-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Sks-devel] Rationalization (Was: Questions regarding blocking ...)


From: Kiss Gabor (Bitman)
Subject: [Sks-devel] Rationalization (Was: Questions regarding blocking ...)
Date: Sun, 3 Aug 2014 09:06:47 +0200 (CEST)
User-agent: Alpine 2.02 (DEB 1266 2009-07-14)

On Sat, 2 Aug 2014, Yaron Minsky wrote:

> I think that increasing the number of gossip peers only improves
> things.  The probability of collisions I think does not systematically
> go up when you increase the number of peers, and it reduces the chance
> that you've chosen an unusually busy set of peers.

Dear folks,

I'm thinking on structure of graph of key servers for a long time.
IMHO it is too scale free and not designed knowingly enough.
There are a few "big" nodes with high vertex number.
(Some key server operators just collect peers like poststamps.)
Other "small" nodes feel enough to connect to a few big ones.
I think we should to force less deviation of degree of nodes.

In parallel I would limit number of my partners on a rational level.
30 or 50 peers do not give more reliability than 20 does.
However they consumes more resources.

So I plan to follow this algorithm:
If a new guy applies for partnership I connect him because
he need much more support initially but the same
time I disconnect from a node with highest degree.
In the beginning while my degree is higher than 20 I drop
the two biggest peers for a new beginner.

What is your opinion?

Cheers

Gabor



reply via email to

[Prev in Thread] Current Thread [Next in Thread]