On Mon, 17 Mar 2003 00:11:26 -0800 Mark J Roberts <mjr at znex.org> wrote:
> I think the performance of the LRU algorithm is an open question,
> by the way.

I had another idea. How about to do some weighting towards smaller keys?

Example:
My node has to delete keys from the data store and there are two 'oldest' keys
key A: last usage: T,         size: 1 KiB
key B: last usage: T+epsilon, size: n KiB (n>1)

Although key A is the oldest I'd rather like to delete key B because it is 
bigger. (Sure, if epsilon isn't an epsilon anymore ;) than I'd delete the key 
A.)

advantages:
- 'widgets' are better available (tiny graphics, smiley-gifs, ...)
- authors are encouraged to compress the data before inserting it

disadvantages:
- authors could use splitfiles of 1k-splits for everything they insert to 
ensure better availability ...

--Palomitas



Concerned about your privacy? Follow this link to get
FREE encrypted email: https://www.hushmail.com/?l=2 

Big $$$ to be made with the HushMail Affiliate Program: 
https://www.hushmail.com/about.php?subloc=affiliate&l=427
_______________________________________________
devl mailing list
devl at freenetproject.org
http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl

Reply via email to