Huge table

[LanceT]

For example, myspace.com has 150 million accounts, how do they store all of it?

[Hughesy1986]

Thye are most likely using a mysql database.

 

Glen

[btherl]

They use SQL server:

 

http://www.baselinemag.com/print_article2/0,1217,a=198614,00.asp

 

There's plenty of other technologies being used to allow them to handle so much traffic.. the database is just one small part.

[LanceT]

They use SQL server:

 

http://www.baselinemag.com/print_article2/0,1217,a=198614,00.asp

 

There's plenty of other technologies being used to allow them to handle so much traffic.. the database is just one small part.

 

SQL can handle large database better than mySQL? Also, can we connect to SQL the same way that we connect to mysql?

 

Is the database structure different? Can we use a .sql file from mySQL and import it into SQL?

[fenway]

SQL Server is a different beast altogether.

[veridicus]

As btherl says, there's far more going on then just a huge database.  You can be sure they're using various levels of caching to prevent too many queries.  With a little searching you can find out if they're using something big like Akamai's services.

 

But on the flip side, I've been lucky enough to work on one of the largest commercial databases in the world.  They inserted 12 to 14 million new records per day into the main table.  With proper queries and cursor scripts I was able to run dynamic reports against it with great speed.  So it is possible to have massive sets of data and good speed if you have the right combination of hardware, database software, DBAs, and developers.

 

And no, it wasn't a mysql database.

[cmgmyr]

http://www.baselinemag.com/print_article2/0,1217,a=198614,00.asp

 

That was a good read.

[LanceT]

As btherl says, there's far more going on then just a huge database.  You can be sure they're using various levels of caching to prevent too many queries.  With a little searching you can find out if they're using something big like Akamai's services.

 

But on the flip side, I've been lucky enough to work on one of the largest commercial databases in the world.  They inserted 12 to 14 million new records per day into the main table.  With proper queries and cursor scripts I was able to run dynamic reports against it with great speed.  So it is possible to have massive sets of data and good speed if you have the right combination of hardware, database software, DBAs, and developers.

 

And no, it wasn't a mysql database.

 

do you think its possible on mySQL database at all? Having good speed that is.

[veridicus]

For really huge tables?  I don't know, but I doubt it.  The big systems give you lots of storage and retrieval options that mysql doesn't.  For example, with Oracle it's often more efficient to walk a data set with a server side cursor than to run a query (the same thing slows down MS SQL Server).  How the data is stored across large or multiple disks also affects query time, so being able to tweak it can make a huge difference.