| From: | "Ciprian Dorin Craciun" <ciprian(dot)craciun(at)gmail(dot)com> |
|---|---|
| To: | "marcin mank" <marcin(dot)mank(at)gmail(dot)com> |
| Cc: | "Michal Szymanski" <dyrex(at)poczta(dot)onet(dot)pl>, pgsql-general(at)postgresql(dot)org |
| Subject: | Re: Using Postgres to store high volume streams of sensor readings |
| Date: | 2008-11-24 05:27:05 |
| Message-ID: | 8e04b5820811232127g6ff6e8balec2145b144e67c9f@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
On Mon, Nov 24, 2008 at 3:42 AM, marcin mank <marcin(dot)mank(at)gmail(dot)com> wrote:
>> Yes, the figures are like this:
>> * average number of raw inserts / second (without any optimization
>> or previous aggregation): #clients (~ 100 thousand) * #sensors (~ 10)
>> / 6seconds = 166 thousand inserts / second...
>
> this is average?
> 166 000 * 20 bytes per record * 86400 seconds per day = 280GB / day ,
> not counting indices.
>
> What is the time span You want to have the data from?
>
> Greetings
> Marcin
Well I'm not sure for the archival period... Maybe a day, maybe a
week... For the moment I'm just struggling with the insert speed.
(We could also use sharding -- horizontal partitioning on
different machines -- and this wourd reduce the load...)
Ciprian.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Thomas Markus | 2008-11-24 07:19:00 | Re: Returning schema name with table name |
| Previous Message | Scott Marlowe | 2008-11-24 02:03:37 | Re: Running postgresql as a VMware ESx client |