Large scale databases - common practice

Hi,
We are designing a large-scale database application. The current one will
receive information from many inputs constantly.
We expect it to grow in one year to 60GB (on average of 4-5GB/Mo).
The database then (after one year) will have a cleaning procedure that will
delete information and the size of the database will not increase
significantly.
Here are a few questions:
1. What are the guide lines to design and implement such a database? Where
can I find information on the web on such cases?
2. From database creation point of view - should we use a few files that
will create the database or should we use only one ? If we use a few files,
what happends to clustered indexes? Where can I find more reading material?
3. The database has the management tables of the application and the data
tables. One data table will hold most of the information in that database. Do
you think I need to divide te table? if so, how do we do it? What is the
common practice in such cases where a huge table holds so much info
(Currently we have a similar database with 27M records in).
how do you build efficient queries? How do you manage the data in the
separate tables? How do you manage indexes? and again where can I find
material on this?
We know to look inside the BOL but we are looking for common practice
solutions.
TIA!!
NoamPlease don't post independently in separate newsgroups. You can add
multiple newsgroups to the header and then all the answers appear as one.
See my reply in the other newsgroup.
Andrew J. Kelly SQL MVP
"Noam" <Noam@.discussions.microsoft.com> wrote in message
news:2DCD3F8D-DFCF-46F7-BB1C-30355D49B321@.microsoft.com...
> Hi,
> We are designing a large-scale database application. The current one will
> receive information from many inputs constantly.
> We expect it to grow in one year to 60GB (on average of 4-5GB/Mo).
> The database then (after one year) will have a cleaning procedure that
will
> delete information and the size of the database will not increase
> significantly.
> Here are a few questions:
> 1. What are the guide lines to design and implement such a database? Where
> can I find information on the web on such cases?
> 2. From database creation point of view - should we use a few files that
> will create the database or should we use only one ? If we use a few
files,
> what happends to clustered indexes? Where can I find more reading
material?
> 3. The database has the management tables of the application and the data
> tables. One data table will hold most of the information in that database.
Do
> you think I need to divide te table? if so, how do we do it? What is the
> common practice in such cases where a huge table holds so much info
> (Currently we have a similar database with 27M records in).
> how do you build efficient queries? How do you manage the data in the
> separate tables? How do you manage indexes? and again where can I find
> material on this?
> We know to look inside the BOL but we are looking for common practice
> solutions.
> TIA!!
> Noam
>|||Thank you for your fast reply. This is the first time I posted here and I
will do as you suggest next time!.
Thanks.
Noam
"Andrew J. Kelly" wrote:
> Please don't post independently in separate newsgroups. You can add
> multiple newsgroups to the header and then all the answers appear as one.
> See my reply in the other newsgroup.
>
> --
> Andrew J. Kelly SQL MVP
>
> "Noam" <Noam@.discussions.microsoft.com> wrote in message
> news:2DCD3F8D-DFCF-46F7-BB1C-30355D49B321@.microsoft.com...
> > Hi,
> >
> > We are designing a large-scale database application. The current one will
> > receive information from many inputs constantly.
> >
> > We expect it to grow in one year to 60GB (on average of 4-5GB/Mo).
> >
> > The database then (after one year) will have a cleaning procedure that
> will
> > delete information and the size of the database will not increase
> > significantly.
> >
> > Here are a few questions:
> > 1. What are the guide lines to design and implement such a database? Where
> > can I find information on the web on such cases?
> > 2. From database creation point of view - should we use a few files that
> > will create the database or should we use only one ? If we use a few
> files,
> > what happends to clustered indexes? Where can I find more reading
> material?
> > 3. The database has the management tables of the application and the data
> > tables. One data table will hold most of the information in that database.
> Do
> > you think I need to divide te table? if so, how do we do it? What is the
> > common practice in such cases where a huge table holds so much info
> > (Currently we have a similar database with 27M records in).
> > how do you build efficient queries? How do you manage the data in the
> > separate tables? How do you manage indexes? and again where can I find
> > material on this?
> >
> > We know to look inside the BOL but we are looking for common practice
> > solutions.
> >
> > TIA!!
> >
> > Noam
> >
>
>