IIS 7，Microsoft Access和Classic ASP的最高效的ADODB连接

Question

I'm managing an Intranet site written in Classic ASP, running in IIS7 on Windows Server 2008, and using Microsoft Access as a datastore (switching to a more capable database is not an option, nor is switching to .NET).

Each page makes multiple calls to the datastore in order to populate various views, etc.

As traffic to the site has grown we are experiencing issues with page response times and one possible cause lies in the efficiency of the connections to the datastore.

What is the most efficient ADO connection type to use for this type of setup?

Answer 1

We had a similar issue with a few of our websites. All the content was database driven. We used SQL Server and made as few database round trips as possible when displaying web pages. We became concerned about heavy usage sites constantly hitting the database and made some changes to how our websites got built going forward.

We wound up replacing about 70% of the database driven content with static include files. We divided the website content data into three groups.

1. Static Content that did not change
2. Semi-Static Content that changed rarely
3. Database Driven Content

We replaced the Static Content with include files. If the customer wanted a change, they told us, we made the change (for a fee).

We classified the Semi-Static content as data that rarely changed but could be changed by the customer using the back-end interface . In this case we also used static include files. After the changes to the database were made, we ran a (transparent to the user) static-data creation widget that recreated the static data as include files based on the database update. The next time the page was called it would use the newly built include file.

Answer 2

I'm not sure you can gain performance from using different ADO types. I think it's more to do with database size and structure.

I had similar troubles to you with my application. Although I have built many, many apps for my clients, not many of them had a huge audience. Recently I have built one, and boy, did I have to learn a thing or two to get it to be efficient.

Here a few things to consider, if you haven't already. Some might be a bit basic for you but useful to others too:

1) INDEX or FULLTEXT INDEX - To speed up my database, I added INDEX's to the columns I query which speeds up my queries a huge amount (but slows down the INSERT's slightly). I also added FULLTEXT INDEX's to the columns where I searched for multiple words, so I could lose the horribly slow LIKE/IN clauses.

2) Normalization or De-normilization - One of my queries searched 16 tables using joins and sub-queries and turned out to be slow, bulky and confusing. I fixed this by creating a "de-normalized" table, that contained all the usable data from those 16 tables that I needed for the search. So with the help of FULLTEXT INDEX and a de-normalized and indexed table, my query is just a few lines in length is remarkably quick. Before using de-normalized tables, you must first understand normalized tables. For me, this was the correct situation where you can break the rules.

3) SQL Field selection - Only select the fields you require in your SQL query. This is a simple mistake but still can be made. For example, don't do (select * from myTable) if you don't need all the columns returned, do something more precise like (select id, fname from myTable), if that's the only columns you need.

4) DB cleanup - I'm not sure about Access but in MYSQL you can cleanup your DB by using the OPTIMIZE query, which repairs any gaps in your DB, sometimes created by inserting and deleting lots of data frequently. You can also use EXPLAIN, which explains your query in detail, such as a timer, indexes used, rows used and records returned.

There are many ways to gain DB performance. If you Google "database optimization" or "database tuning", you should find some good reads..