sks-devel
[Top][All Lists]
Advanced

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

Re: [Fwd: Re: [Sks-devel] Alternative to round-robin (was Re: pool.sks-k


From: David Shaw
Subject: Re: [Fwd: Re: [Sks-devel] Alternative to round-robin (was Re: pool.sks-keyservers.net having trouble?)]
Date: Tue, 6 Apr 2010 14:37:50 -0400

On Apr 6, 2010, at 2:28 PM, Jeff Johnson wrote:

> 
> On Apr 6, 2010, at 1:43 PM, David Shaw wrote:
> 
>> On Apr 5, 2010, at 5:01 PM, Kristian Fiskerstrand wrote:
>> 
>>> I also did some minor fiddeling around with the priorities. It still
>>> only returns a limited amount of records (currently 10), but those with
>>> a higher weight are always prioritized in the selection as well.
>>> 
>>> Feel free to come with input as to the number of records to show and
>>> which servers should benefit from weighting up. Currently I set my own
>>> keys.kfwebs.net to weight 15 and keyserver.gingerbear.net. to weight 30
>>> for test purposes. The default for the rest is 20.
>> 
>> There are many ways to automate this if you wanted to.  One simple way to do 
>> it would be to rank the servers on how fast they respond.  Faster servers 
>> will thus be queried more, until such time as they are not fast, and then 
>> they'd be queried less.  Slower servers would be queried less to allow them 
>> to recover and become fast.  It's somewhat naive, but should match the 
>> real-world performance better than the round-robin does.
>> 
> 
> The "preference" algorithm is far easier to state than to
> implement, mostly because whatever server _WAS_ fastest
> is rather different than whatever server _IS_ fastest.

Just rig it up similarly to how DNS does it.  Rank things in order of 
performance, and then artificially penalize the fastest ones on each iteration 
until they aren't the fastest any more.  Then repeat.  The goal isn't 
perfection - it's just to be better than round robin :)

David



reply via email to

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