|
ASP Performance
Part:
1
2
(Continued from previous part...)
Response Time of "Hello" Page
The next area I want test is total response time of ASP pages. To do this, I wrote
the following Java program. This program is doing a single HTTP request to
a specifield Web page, and repeating this for many times.
/**
* HttpResponseTest.java
* Copyright (c) 2002 by Dr. Herong Yang. All rights reserved.
*/
import java.io.*;
import java.net.*;
public class HttpResponseTest {
public static void main(String[] args) {
int numberOfTests = 1;
if (args.length > 0) numberOfTests
= Integer.valueOf(args[0]).intValue();
long t1 = System.currentTimeMillis();
String result = "";
for (int nTest=1; nTest<=numberOfTests; nTest++) {
result = test(args);
}
long t2 = System.currentTimeMillis();
long t = t2 - t1;
PrintStream out = System.out;
out.println("Performace Information:");
out.println(" Number of tests = " + numberOfTests);
out.println(" Time = " + (t/1000) + " seconds.");
out.println("Rerulst of Last Test:");
out.println(result);
}
public static String test(String[] args) {
String path = "/index.html";
int port = 80;
String host = "localhost";
if (args.length > 1) path = args[1];
if (args.length > 2) port
= Integer.valueOf(args[2]).intValue();
if (args.length > 3) host = args[3];
String result = "";
try {
Socket c = new Socket(host,port);
BufferedWriter w = new BufferedWriter(new OutputStreamWriter(
c.getOutputStream()));
BufferedReader r = new BufferedReader(new InputStreamReader(
c.getInputStream()));
String m = "GET "+path;
w.write(m,0,m.length());
w.newLine();
w.flush();
while ((m=r.readLine())!= null) {
result = result + m + "\n";
}
w.close();
r.close();
c.close();
} catch (IOException e) {
System.err.println(e.toString());
}
return result;
}
}
Now compile this program and try it with the following "Hello" ASP page, hello.asp:
<%@ language="vbscript"%>
<html><body>
<%
response.write("Hello world!")
%>
</body></html>
you will get something similar to this:
\local\j2sdk1.4.1_01\bin\java -cp . HttpResponseTest 1 /asp/hello.asp
Performace Information:
Number of tests = 1
Time = 0 seconds.
Rerulst of Last Test:
HTTP/1.1 200 OK
Server: Microsoft-IIS/5.0
Date: Tue, 24 Dec 2002 02:47:31 GMT
Connection: Keep-Alive
Content-Length: 38
Content-Type: text/html
Set-Cookie: ASPSESSIONIDGGGQQWLY=BJINNIGBFMBPMNHNNCJFDJLG; path=/
Cache-control: private
<html><body>Hello world!</body></html>
I repeated the tests by changes the controlling parameters.
The following table shows the results comparing with similar tests I did
with other technologies:
Number Debug Time
Cases of Tests Mode (Sec) Note
1. 1000 No 2 Static text with IIS 5.0
2. 2000 No 4 Static text with IIS 5.0
3. 1000 No 6 ASP page with IIS 5.0
4. 2000 No 11 ASP page with IIS 5.0
5. 1000 ? 7 Static text with Tomcat 4.1.18
6. 2000 ? 15 Static text with Tomcat 4.1.18
7. 1000 ? 8 JSP page with Tomcat 4.1.18
8. 2000 ? 16 JSP page with Tomcat 4.1.18
Conclusion:
- So the performance cost of writing text through ASP statements is
2 times slower than the static pages.
- Tomcat is 3 times slower than IIS when serving static pages.
- Tomcat is also slower than IIS when serving dynamic text.
Part:
1
2
| 