On Apr 15, 2009, at 6:11 PM, Michael Orlitzky wrote: > 2) The bandwidth overhead of SSL is negligible. Seriously.
A quote from an article about using SSL - <http://searchsoa.techtarget.com/news/interview/ 0,289202,sid26_gci995388,00.html> " The increase in message size due to SSL is not very significant, and is rarely a concern. " It isn't a bandwidth issue, it is a data throughput issue at the server and at the client. The server has to calculate the encryption, and that takes time, so the response from a server using SSL / TLS will lag compared to unencrypted traffic. Likewise the browser has to decrypt the data, so there is a lag while the local CPU does the calculations before it can parse the HTML / JavaScript. The length of those lags is dependent on the processing power at each end, not dependent on bandwidth. Here <http://www.webperformanceinc.com/web_stress_test/ performancerealistic.html#CPU> you can see that a 2.4 GHz server can deliver better performance serving encrypted traffic than a 800 MHz server serving unencrypted traffic. Yeah, it's comparing a Fiat 500 to an Alfa-Romeo, but the numbers comparing encrypted vs, unencrypted for each CPU are interesting. Although I am not that familiar with satellite links, I do know that one of the biggest problems using satellite is latency. The lag at the server while the encryption is being calculated is the pain point; it adds to the latency problem. > I don't mean to be rude, but this is a bad idea, I think it isn't such a good idea either. If you have e-mail messages that fall under a NDA, you'd want that connection encrypted at all times, not just to protect the authentication. If you have the time and the skill to hack on the code, this might be doable, but what you are looking for isn't built in standard. -- Charles Dostale System Admin - Silver Oaks Communications http://www.silveroaks.com/ 824 17th Street, Moline IL 61265 _______________________________________________ List info: http://lists.roundcube.net/users/
