My guess is the possible slowdown may be related to the order the security protocols are polled. Depending on how the server challenges the responses from the client may be different and result in multiple 401 requests to the server to get the right auth mode.

Ultimately the way to check is run Fiddler traces on both and see what the actual traffic difference is.

I"ve never found HTTP client to be a performance issue, but then i've never really checked either. But I run several applications that pull large chunks of data almost instantly so that it never really mattered.

+++ Rick ---

>I also think the time is unusually long, but in non-HTTPS areas sometimes convoluted login+encrypt scenarios were/are used to harden security.
>My first attempt would be to measure each distinct action to see if login is responsible for most of the duration
>
>>Hi Thomas,
>>
>>Until Rick Strahl shimes in (quite sure he will), 2-3 seconds is just abnormal for what you describe -- should be a hundred times less (20-30 ms).
>>
>>What does your VFP server do exactly?
>>Do you use west-wind web connect on your server?
>>
>>>Hi there,
>>>
>>>as I am in the process of planning a new VFP-application I came across a problem I noticed when I compared two apps one written in vb.net and one in vfp using the wwipstuff-tools. In both apps I have implemented a http-based REST-Service which consists of a login, a request and a logout. So far, so good. But what I have noticed is that the VFP-App is slower than the vb.net-app.
>>>The first one needs 2-3 seconde, the second one 1 second. At first this sounds not to important but if it is used a hundred times or more it becomes important.
>>>What I would like to ask you is, if there ore other methods of implementing this via VFP, which are not wininet-based?
>>>I never used the MSXMLHHTP-stuff, does anyone know if it is faster? Or would it be better to use the .net-functionality in a com-server?
>>>
>>>Best regards and thanks in advance
>>>
>>>Thomas