Re: Synchronization routine
- From: "Joris Dobbelsteen" <joris.dobbelsteen@xxxxxxxx>
- Date: Mon, 31 Oct 2005 12:35:34 +0100
For some funny reason I see your text as a attachment in outlook express...
"Ertugrul Soeylemez" <never@xxxxxxxxxxxxxx> wrote in message
news:20051031043743.0328ed08@xxxxxxxxxxxxxxx
>"Joris Dobbelsteen" <joris.dobbelsteen@xxxxxxxx> (Sun, 30 Oct 2005 21:30:48
>+0100):
>> Basically I have two lists, one on my firewall and one in a database. The
>> list on the firewall must become the same as the list in the database.
>> The
>> current sizes are over 1 million entries and I expect to see 2 million
>> rows
>> in the future.
>
>Taking into account that you have an old machine, you might be very low
>on memory, so this isn't practical. For about one million entries, the
>memory consumtion is the same as the size of the individual entries, in
>MB! If every entry is 128 bytes in size, then a million entries will
>consume roughly 128 MB. And the data size is well under-estimated. You
>should consider this.
I did consider this.
The average entry size is 24 bytes. So this makes 24 MB for 1 M entries.
With pointers and such it will grow probably several MBs too, but not that
far. The data from the database was pulled for a sequential scan only, not
to comsume too much memory.
The previous (probably flawed) implementation used over 128 MB for only 40k
entries. The dataset was very similair, so I would have expected only
several MB, not growing fast too. Probably something here went wrong.
- Joris
.
- References:
- Synchronization routine
- From: Joris Dobbelsteen
- Re: Synchronization routine
- From: Ertugrul Soeylemez
- Synchronization routine
- Prev by Date: Re: Programming: Not stressful? Yeah, right...
- Next by Date: Re: please check my homework
- Previous by thread: Re: Synchronization routine
- Next by thread: bug in windows getsavefilename api?
- Index(es):
Relevant Pages
|