大规模数据库性能…MongoDB / NoSQL eval

Question

I'm working on a project currently where in production we have 100k rows or so in production.

The existing system which I inherited is having quite a bit of problems dealing with this much data due to the complexity of the application. This will only be compounded year over year with the about the same frequency. Give we are already having problems I think I need to start investigating options.

This was written in .NET / MSSQL. I'm trying to evaluate and determine if using a different database structure like MongoDB or maybe even mixed would help improve this. For example, if the data was modeled in json it would look something like:

Record

 {
    1234: "Harry",
    1233: "Black",
    1234: 5
 }

Meta

{
    1234: {
        name: "First Name",
        alias: "fname",
        guid: "343-3433-4333",
        initalValue: "Calvin",
        type: 1
    },
    1233:{
        name: "Last Name",
        alias: "lname",
        guid: "343-3433-4333",
        initalValue: "Marshall",
        type: 1
    },
    1233:{
        name: "Age",
        alias: "age",
        guid: "343-3433-4343",
        type: 2,
        listId: 1
    }
}

List Meta

{
    1:{
        1: "0-20",
        2: "20-30",
        3: "30-50",
        4: "50-60",
        5: "60 and older"
    }
}

This is a basic example, there is much more data in the record and more complexity in the meta data portion. Given this, what type of system would you suggest?

Answer 1

If you use SQL Server 2012, you can consider ColumnStore Index technology : http://msdn.microsoft.com/en-us/library/gg492088.aspx , especially the scenarios : http://msdn.microsoft.com/en-us/library/gg492088.aspx#Scenarios , and check if it can fit.

This tehcnology will be improved in the new version of SQL Server, SQL Server 2014.