C# Licensing System

That way it looks like all traffic is coming from the client IP, not your server. Your bot might not be so effective if everyone's traffic comes from a single source. You could send your own proxy address for the server to connect to, but people might not be so happy about their proxies being harvested (of course whatever other data you send can be harvested too).

You can also selectively decide what the client proxies and what it doesn't (less bandwidth issues for you when transferring non sensitive data), they'll share the same IP as far as the target server is concerned no matter where the request started, and you can share cookies between your server and client easily enough, all that will change is the SSL client's public key which is just random anyway (you could make your server key permanent if you so desired).

Also if it's only the request formation, not the response and post-processing, that's critically important then your server can just send the encrypted request ready for you to forward with the clients on public key so they can read the response, but not the request.

ampedsoftware said:

That way it looks like all traffic is coming from the client IP, not your server. Your bot might not be so effective if everyone's traffic comes from a single source.

Click to expand...

All non-trivial bots require proxies anyway.

You could send your own proxy address for the server to connect to, but people might not be so happy about their proxies being harvested (of course whatever other data you send can be harvested too).

Click to expand...

That 's true but when someone uses a proxy on a third party program, he can't know if it 's harvested or not anyway whether the software uses a server or not.

That 's true but when someone uses a proxy on a third party program, he can't know if it 's harvested or not anyway whether the software uses a server or not.

Click to expand...

Not true for HTTPS, unless your bot uses a shitty SSL library that's easy to trick or the proxy owner happens to have an illegitimate server cert (which is certainly possible, but unlikely for any given single proxy/cert combo). Unless you're talking about the software harvesting the proxy address/port and not the data, which you could probably tell from Wireshark; granted only 0.01% of the population is going to check this, but its possible.

All non-trivial bots require proxies anyway.

Click to expand...

Depends, if you're doing some crawling and you don't need to hit a server many times very quickly, then you don't need them. Regardless, by using the server to connect to things directly (even through provided proxies) you could be in for some headaches if people start using your bot for illegal purposes and the path traces back to you first. You could keep logs of user traffic so if someone knocks on your door you can show that it wasn't you, it was the client, but that's another hassle you don't need. And by keeping server bandwidth to a minimum through tunnelling only what you need you can save a bit on your hosting costs. Many bots could cripple a good server if the server had to pipe every bot user's traffic.

ampedsoftware said:

Unless you're talking about the software harvesting the proxy address/port and not the data

Click to expand...

^^ this

How about this for an idea.

Take a section of code, upload it to your server.

.NET allows runtime compilation of text string into a a runnable object.

Users must enter a valid username/password and then the string code gets sent for compilation and running.