DEV 411 ASP NET Best Practices For Performance

DEV 411 ASP. NET: Best Practices For Performance Stephen Walther www. Superexpert. Training. com

Purpose of Talk What is the fastest method of displaying a set of database records in an ASP. NET Page?

Testing Tools Trace Tools Profiler Tools Load Tools

Trace Tools ASP. NET Page or Application Tracing Display trace information on page System. Diagnostics Tracing Write trace information to custom listener

Trace Tools

Profiler Tools CLR Profiler Free Profiler from Microsoft ANTS Profiler Available at www. Red-Gate. com SQL Profiler Included with Microsoft SQL Server

ANTS Profiler

Load Tools Application Center Test Included with Visual Studio. NET Enterprise Web Application Stress Tool (WAST) Free download from Microsoft site ANTS Load Available at www. Red-Gate. com. Includes calculated Frustration Coefficient

Test Setup Application Center Test (ACT). NET Framework 1. 1 SQL Server 2000 Windows Server 2003 (Standard)

Performance Statistics Requests Per Second (RPS) Time To Last Byte (TTLB) Page Execution Time

Timer Module The Timer Module records the time interval between Pre. Request. Handler. Execute Event Post. Request. Handler. Execute Event

Begin. Request Pre. Request. Event. Handler. Execute Page Events Application Events Timer Module Init Load Unload Post. Request. Event. Handler. Execute End. Request Timer. Module. cs

Clock Resolution Query. Performance. Counter is accurate to 1/3579545 of a second or about a millionth of a second

The Test Request page 1050 times Discard first 50 requests Log time of each request Average results

Database Setup Four Database Tables Products 10 – 10 Rows Products 50 – 50 Rows Products 100 – 100 Rows Products 500 – 500 Rows

What’s Faster? Data. Reader Data. Set Display. Data. Reader. aspx Display. Data. Set. aspx

Data. Reader

Data. Set

Data. Reader Versus Data. Set

Data. Reader Versus Data. Set Final Results On average, a Data. Reader is 16% faster than Data. Set

3 rd Option – Array. List Using an Array. List instead of a Data. Reader results in similar performance with the advantages of a static representation of data Display. Array. List. aspx

Array. List

What’s Faster? Sql. Data. Reader Ole. Db. Data. Reader

Ole. Db. Data. Reader

Ole. Db. Data. Reader Final Results On average, a Sql. Data. Reader is 115% faster than an Ole. Db. Data. Reader

What’s Faster? Inline SQL Stored Procedure

Stored Procedure

What’s Faster? Data. Reader Column Reference By Name: Response. Write(dr[“Product. Name”]); By Ordinal: Response. Write(dr[0]); By Get. String(): Response. Write(dr. Get. String(0));

Column Reference

Column Reference Final Results On average, ordinal reference is 11% faster than by name

What’s Faster? Proper Case dr[“Product. Name”] Improper Case dr[“PRODUCTNAME”]

Proper Case

Proper Case Final Results Using proper column case is 1% faster than improper column case

What’s Faster? Inline ASP. NET Controls

Data. Grid

Data. Grid Final Results Inline script is 233% faster than a Data. Grid

What’s Faster? Data. Grid with View. State Disabled Data. Grid with View. State Enabled

View. State

View. State Final Results Data. Grid with View. State disabled is 66% faster than Data. Grid with View. State enabled

What’s Faster? Auto. Generate. Columns Template Columns

Template Columns

Template Columns Final Results A Data. Grid without templates is 39% faster than a Data. Grid with templates

What’s Faster? How to improve template performance? Data. Binder. Eval <%# Data. Binder. Eval(Container. Data. Item, “Product. Name”) %> Explicit Cast <%# ((Db. Data. Record)Container. Data. Item)["Product. Name"]%> Item. Data. Bound void Item. Data. Bound(Object s, Data. Grid. Item. Event. Args e) Display. Item. Data. Bound. aspx

Template Performance

Template Performance Final Results Explicit cast is 11% faster than using a databinding expression

Creating A Custom Control Would a custom Data. Grid (with severely reduced functionality) be faster than the standard Data. Grid? Fast. Grid. cs

Custom Control

Custom Control Final Results Fast. Grid is 37% faster than a standard Data. Grid

What’s Faster? Data. Grid with no caching Data. Grid with data caching Data. Grid with output caching

Data Caching

Data Cache Final Results Using the data cache is 637% faster than a standard Data. Grid

Output Cache

Output Cache Final Results Using the output cache is infinitely faster than using a standard Data. Grid

Conclusions A Data. Reader is faster than a Data. Set An inline Data. Reader is faster than a Data. Grid You pay a high price for View. State Auto. Generate. Columns is faster than template columns Caching is always a good idea!

Performance Resources Microsoft Sites Performance http: //msdn. microsoft. com/netframework/using/understanding/perf/default. aspx Books Improving. NET Application Performance and Scalability http: //msdn. microsoft. com/library/en-us/dnpag/html/scalenet. asp Software Web Application Stress Tool http: //www. microsoft. com/technet/itsolutions/intranet/downloads/webstres. mspx CLR Profiler http: //www. microsoft. com/downloads/details. aspx? Family. ID=86 ce 6052 -d 7 f 4 -4 aeb-9 b 7 a 94635 beebdda&displaylang=en ANTS Profiler and ANTS Load http: //www. Red-Gate. com

Please fill out a session evaluation on Comm. Net Q 1: Overall satisfaction with the session Q 2: Usefulness of the information Q 3: Presenter’s knowledge of the subject Q 4: Presenter’s presentation skills Q 5: Effectiveness of the presentation