Dear support team,
My team is developing an Internet application that will be used by people located in different countries and having different quality of Internet access (bandwidth, latency...).
I would like to test this application by emulating the different types of network situations our users will face.
Connection emulator looks very good but, before buying, I would like to get a few questions answered:
1) I would like to run Connection Emulator on the server running the application. Will I need multiple network interfaces on this server, or can Connection Emulator act as a TCP or UDP proxy using one single Network interface on different ports? Any documentation on how Connection Emulator acts on the network would be welcome.
2) How could I "measure" the network characteristics of a given user, to define what parameters values I should use to emulate this network with Connection emulator?
3) With the trial version, in the "Filter settings", I didn't manage to activate the Source and Destination filters, even by checking the 'enable this filter" checkbox. Are these features inhibited in the trial version or did I do something wrong?
Thank you!
Benoit
Questions on Connection Emulator before buying
Started by Benoit Doreau
|
|
Questions on ConnectionEmulator before buying 16 November 2012, 11:57 |
Hi Benoit,
1. It depends where and how you would to simulate those parameters. Basically you can think of CE as of a firewall or a filter that sits between a network card and the TCP/IP stack, inserting delays and dropping packets as necessary.
Since you are developing an Internet application, assuming that it runs on a Windows server and clients access it via a web-browser, you can either install CE on the server or on one or more clients (provided they are running Windows). You can easily try it out, by installing CE on your computer and ping any local address. Then start a simulation and add latency, you will see ping response times increasing.
2. I am not so sure, how a physical line's characteristics can be measured, but I'd suggest to run a ping, say 10000 requests, and upon completion it will give you an average latency and packet loss. As in example below:
Also you can run a speed test to determine maximum upload and download rates.
Finally you specify some or all of these parameters: bandwidth limit, packet loss, latency in CE.
3. You just need to choose an IP protocol, e.g. TCP or UDP to activate the source and destination filters. This is because when Any protocol is chosen, it really means Any protocol, including non-IP, which don't have IP addresses.
Hope this helps
1. It depends where and how you would to simulate those parameters. Basically you can think of CE as of a firewall or a filter that sits between a network card and the TCP/IP stack, inserting delays and dropping packets as necessary.
Since you are developing an Internet application, assuming that it runs on a Windows server and clients access it via a web-browser, you can either install CE on the server or on one or more clients (provided they are running Windows). You can easily try it out, by installing CE on your computer and ping any local address. Then start a simulation and add latency, you will see ping response times increasing.
2. I am not so sure, how a physical line's characteristics can be measured, but I'd suggest to run a ping, say 10000 requests, and upon completion it will give you an average latency and packet loss. As in example below:
Also you can run a speed test to determine maximum upload and download rates.
Finally you specify some or all of these parameters: bandwidth limit, packet loss, latency in CE.
3. You just need to choose an IP protocol, e.g. TCP or UDP to activate the source and destination filters. This is because when Any protocol is chosen, it really means Any protocol, including non-IP, which don't have IP addresses.
Hope this helps