Scale Out: Distributed Caching On The Path To Scalability

SOA makes application scalability easier, but getting the best performance out of your SOA means paying close attention to the details. We'll give you some tips for coding your app more effectively and using distributed caching to remove data access bottlenecks.

Iqbal Khan

MSDN Magazine June 2010

This column covers Web application performance with distributed caching, a utility for object-object mapping, and a look at interesting blogs.

Scott Mitchell

MSDN Magazine September 2009

Microsoft Velocity exposes a unified, distributed memory cache for client application consumption. We show you how to add Velocity to your data-driven apps.

Aaron Dunnington

MSDN Magazine June 2009

Currently I'm using SSIS packages, C#,Powershell, BulkInserts technique for ETL and I’ve one database with different schemas.
for example core.table1, core.table2, trans.table1, trans.table2 in SQL-2008.

I would like to create 3 different SQL physical nodes for load balancing, performance and to handle large scale of data.

Two simple questions:
1) create same table structure in all 3 node, and load data parallel,
 for example if I’m receiving 1000 records- distribute 1000 records in 3 nodes (300 + 300 + 400),
 To retrieve the data I should have same algorithms, three separate thread process parallels and get me the result form 3 nodes.

2) Create one schema on one node, another in second node and dump the data accordingly.
       for example all core. tables are under node1, trans. tables are under node2 and so on..

please let me know the pros & cons of above approaches?

Is that Store procedure adviseable in distributed SQL environmnet?
I feel LINQ-SQL is more convineant as application get the data by using DBML Schema? Please advise

Which would be preferable approch on ETL,
1) Extract > Transform > Load
2) Extract > Load (Flat file Load) > Trnasform > Load

 I'm conf

CAB from MS is a very famous tool for the developers as it saves a lot of valuable time of the developers. But anyone who has decided to use CAB, should also keep into mind its limitations as well. The cache in EnltLib is stand alone and in-process in nature that’s why it can ends up with some scalability and reliability issues when used in a multi-server environment. But there are some party integrations which can over come this issue. The most popular way to overcome these limitations of CAB is the use of a distributed cache. The article about the drawbacks of CAB and it solutions can be a very helpful read in this regard. 

 I get the following error on last line below (working on remote host): 'E:\web\sfcascom\ls_admin/uploads/images/001.JPG' is not a valid virtual path,

Note that the slashes go in different directions

I've tried a number of things but nothing seems to work. Thanks-

Imports System.Drawing
Imports System.Drawing.Drawing2D
Imports System.Drawing.Imaging
Imports System.Web.UI.WebControls

Imports System.Drawing.Imaging.ImageCodecInfo
Partial Class Protected_imageupload
    Inherits System.Web.UI.Page
    Private Sub btnUpload_Click(ByVal sender As Object, ByVal e As System.EventArgs) Handles btnUpload.Click
        ' Initialize variables
        Dim sSavePath As String
        Dim sThumbExtension As String
        Dim intThumbWidth As Integer
        Dim intThumbHeight As Integer
        Dim sFilename As String = ""
        ' Set constant values
        'sSavePath = "images/"
        sSavePath = Server.MapPath("~/") + "ls_admin/uploads/images/"



        sThumbExtension = "_thumb"
        intThumbWidth = 160
        intThumbHeight = 120

        ' If file field isn?t empty
        If FilUpload.PostedFile IsNot Nothing Then
            ' Check file size (mustn?t be 0)
            Dim myFile As Http

I'm having a problem with an asp.net 3.0 application using a gridview to display data.  The problem seems to be that despite a brand new .selectCommand being sent to the sqlDataSource, the gridview starts displaying data from *someone else's* query.  For example, user A is logged in and will be working fine until a few page submissions later when the gridview data starts suddenly showing information that belongs to another user. 

This is my first asp.net application, though I've been working with classic asp and ADO for years so I'm used to sending a SQL string and getting back the results from that string.  Apparently there is a complex caching system which is interferring which I do not at all understand so at this point I just want to stop the server from caching any data whatsoever.  Is this possible?

Any and all help will be much appreciated!!

I am just trying to understand the basics flow to the MVC framework by putting in some breakpoints in the beginning of the actions of a couple of controllers.

One break point is at the HomeController/Index action where it just returns View() of the "Welcome to MVC".  The Second break point is at a SecondController/Index action where it uses the Entity Framework and returns a View() of a list from the db.

When I run this in debug, it breaks at the points I expected when I go from Home page to the Second page controller and back to Home.  However, when I go to the Second page again, it doesn't break at all?  Is the results of the Second page being cached and how and where?  How do you trigger for it to get check if data is old?

I check to see if the persistence is done on the db, by changing the value of a field on the db side and it still did not cause it to break on that action and displays old information in the view.  I am concerned because it is displaying incorrect values on the UI.  I do not have the OutputCache attribute above this action or specified any cache on my web.config.

Thanks in advance for your time :)