Cache Storage for Recordset :: Article - 2eNetWorX

Improving Database Performance by Caching serialized Recordsets

by Hakan Eskici

Cache Storage

Having proposed the interface of our Cache class, we now make use of the global.asa file for our internal storage.

We will use an application level Scripting.Dictionary object which is very stable and thread-safe. It enables us to store key-value pairs, and to retrieve stored values by specifying keys. In our implementation, we will use the SQL statements as the unique key.

However, this would impose a risk of messing things up when the same SQL strings are present for different databases. For example, if there are two databases with Customers tables in each, our cache engine might not know which Customers table you are referring to. It is possible to overcome this issue by prefixing the SQL statement with the connection ID (or maybe the connection string itself) to generate really unique keys.

If your website already has a global.asa file, place the following line on the top of it:

<Object Runat="Server" Scope="application" ID="dDataCache" 
ProgID="Scripting.Dictionary"></Object>

Implementation

We have a storage, and a class interface. Now it is time to implement the features.

Let's start with the constructor and destructor of our class. A constructor is a special method of a class that is automatically executed when the class is instantiated by declaring it as an object. A destructor is the method that is run when the object is disposed (destroyed by setting to Nothing).

Class DataCache

'Connection string to connect to the database
Public ConnectionString

'Stream object for our internal use
Private stCache

'Constructor
Private Sub Class_Initialize()
	Set stCache = Server.CreateObject("AdoDB.Stream")
	stCache.Type = 1 'adTypeBinary
end sub

'Destructor
Private Sub Class_Terminate()
	Set stCache = Nothing
End Sub
	
End Class

We use the AdoDB.Stream object for serialization & deserialization of the Recordsets. Therefore we create one in the constructor.

As being the core of our cache engine, now we will build the mighty GetRecordset() function.

< Previous Part: Key Features and Design
> Next Part: GetRecordset Function