Inside and Out...

An attempt to understand technology better...

  Home  |   Contact  |   Syndication    |   Login
  160 Posts | 0 Stories | 12 Comments | 181 Trackbacks

News


WinToolZone - Spelunking Microsoft Technologies
I work as a developer on the Common Language Runtime (CLR) team, specifically in the areas of exception handling and CLR hosting.
Disclaimer

The information in this weblog is provided "AS IS" with no warranties, and confers no rights. This weblog does not represent the thoughts, intentions, plans or strategies of my employer. It is solely my opinion. Inappropriate comments will be deleted at the authors discretion. All code samples are provided "AS IS" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose.

Twitter





Tag Cloud


Archives

Post Categories

Image Galleries

Links

One of the frequently asked questions at the India Windows Communication Foundation Tour was how does WCF scale as compared to existing technologies, like ASMX? Will WCF services be more scalable than the ones we author today? What do we have to do for better scale?

I thought of doing some very basic performance tests to get some data for these questions. The test was done to compare how ASP.NET 2.0 WebServices perform against WCF services. I did these tests on my machine which has 2GB RAM, 60GB HDD and running Windows XP SP2 and IIS 5.1. They were done against NOV CTP of WinFX, building the WCF service using VS 2005 RTM under the debug mode. Below are the results of the same.

Disclaimer: These results are not, by any means, definitive and are just the output of the tests I did. WCF continues to evolve and by the time it is released, the results could vary. You may have different results.

The following are the properties of the tests:

  • Entities tested were ASP.NET 2.0 WebService with BasicProfile 1.1 WebServiceBinding on IIS 5.1 and the WCF self-hosted console executable service (debug mode) with BasicHttpBinding enabled.
  • Three types of load tests were created using VSTS
    • Starting with 10 user load till maximum 250 users, incrementing users by 50 every 10 seconds
    • Starting with 10 user load till maximum 1000 users, incrementing users by 50 every 10 seconds
    • Constant load of 500 users
  • Test ASP.NET 2.0 webservice can be download from here, and Test WCF service can be downloaded from here.

Starting with 10 user load till maximum 250 users, incrementing users by 50 every 10 seconds

ASP.NET 2.0 WebService:

WCF Service:

Result: WCF handles 51.2% more average test cases per second than ASMX with 4.38 times faster average test time.

Starting with 10 user load till maximum 1000 users, incrementing users by 50 every 10 seconds

ASP.NET 2.0 WebService:

WCF Service:

Result: WCF handles 56.79% more average test cases per second than ASMX with 18.82 times faster average test time.

Constant load with 500 users

ASP.NET 2.0 WebService:

WCF Service:

Result: WCF handles 124.11% more average test cases per second than ASMX.

More tests to follow...

posted on Friday, January 06, 2006 6:50 PM

Feedback

# WCF Performance 1/6/2006 12:56 PM On the road to Indigo
One of the most frequently asked questions we're asked about is how WCF performs and scales relative...

# re: ASP.NET 2.0 WebService vs WCF Nov CTP Performance Comparison 1/6/2006 11:24 PM John Ingres
Impressive. Would be very interesting to also compare ES and Remoting since the main reason to use either of them is performance.

# re: ASP.NET 2.0 WebService vs WCF Nov CTP Performance Comparison 1/9/2006 4:43 AM MaheshKumar.R
Great effort, Simple , neat and Clear to anyone.
thanks.
Mahesh

# re: ASP.NET 2.0 WebService vs WCF Nov CTP Performance Comparison 2/7/2006 2:18 PM Mike
Interesting, however in the first test the two scenarios are not the same; you can see that the user ramp-up is not identical.

# re: ASP.NET 2.0 WebService vs WCF Nov CTP Performance Comparison 11/30/2006 10:22 AM Ashish Khadse
Good! I can not get why the speed is better in WCF as compared to Web Service.

# re: ASP.NET 2.0 WebService vs WCF Nov CTP Performance Comparison 1/12/2007 10:27 PM Ken
What tool did you use for testing?

# re: ASP.NET 2.0 WebService vs WCF Nov CTP Performance Comparison 2/14/2007 1:26 AM JackYang
As you know, WebService is the slowest among all of available remoting technologies. Comparing with the slowest web service, it does not point out the WCF performance at all. It will be very interesting if you compare ES, Remoting, DCOM and socket with WCF since the main reason to use them is performance and scalability.




# re: ASP.NET 2.0 WebService vs WCF Nov CTP Performance Comparison 3/14/2007 9:53 PM Paritosh M
Good results but i agree with JackYang. A more apt and useful comparison would be with Remoting

# re: ASP.NET 2.0 WebService vs WCF Nov CTP Performance Comparison 5/21/2007 4:51 AM Jacob
I agree with Jackyang... The heavy HTTP layer isn't used while testing WCF.
It could obviously change the results.


# WCF is actually slow in comparison to raw socket 7/12/2007 11:05 AM JackYang
This site http://www.udaparts.com/document/articles/fastsocketpro.htm

Comments have been closed on this topic.