Running a megasite | SQL Server Performance Forums

SQL Server Performance Forum – Threads Archive

Running a megasite

Hello! I came across a session called "running a megasite on microsoft technologies", which focues on the scalability issues that comes with a large site. More specific about myspace.com and microsoft.com. You can find it here if you missed it:http://sessions.mix06.com/ They are among other things talking about partioning data. And they put 1 million users in one database, and the next upcoming 1 million in another database.
Now when they have several hundreds of millions users, that means several hundreds databases, just to handle the table with users. Then of course they have databases for forum, blog, favourites, and more. In the session they are talking about 650 databases, and growing. This sounds fine, but how do they work with all this data? Lets say that you want to find all users in a specific age range. How is that done when all the data is spread out?
Do they link the servers together using the "linked servers"-feature in SQL-Server?
Do they fire x-hundred asynced queries and mash the result?
If they are using linked servers, are they using a view to query against? How are they working with the tables/databases? Take a mail-database for example. What kind of information is stored there? Do they store just the mail, and then join against the user tables (view?), or do they store duplicated data in the mail-database, for example the username or city who wrote the mail and that sort of thing? If they store all that data duplicated. Do they even use joins? How do they manage to update all this stuff when some user change a setting? Lets say the user change city. You would need to update in many many places where the data is redundant. Maybe i am off the road totally with my statements/questions, if so, please tell me some more about how this is done. Thanks!
If it has been built by and within Microsoft then why don’t you contact them directly, its better to take a direct route.
Once I know further details I will post it here. Satya SKJ
Microsoft SQL Server MVP
Writer, Contributing Editor & Moderator
http://www.SQL-Server-Performance.Com
This posting is provided AS IS with no rights for the sake of knowledge sharing. The greatest discovery of my generation is that a human being can alter his life by altering his attitudes of mind.
]]>