Ed Tomlinson <[EMAIL PROTECTED]> writes:
>On September 22, 2003 02:12 pm, Benny Amorsen wrote:
>> On 2003-09-22 at 19:18, Ian Clarke wrote:
>> > Terrible idea - specialization should occur naturally, it doesn't
>> > have to be forced artificially, and doing so will be harmful.
>>
>> Right now specialization only occurs when nodes are filling up
>> their allocated storage, right?
>>
>> If so, I think it will fail to happen for a /lot/ of nodes. When
>> you have cheap disks and expensive bandwidth, it takes just about
>> forever to fill up a hard drive. Bandwidth should be allocated to
>> where it does the most good, which is to the queries a node
>> specializes in. The same goes for CPU.
>
>A node really starts to specialize when its datastore is nearly full.
>Before this is will save all data it sees over the thresshold it will
>start skipping data that is far from its specialization.

No, specialization occurs because you only get sent certain keys, not
because you select which keys to keep.  Your datastore can be nearly
empty and still be specialized if all the keys it does have and close
together.

theo


_______________________________________________
Devl mailing list
[EMAIL PROTECTED]
http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to