- From: Salvatore Loreto <salvatore.loreto@ericsson.com>
- Date: Mon, 3 Nov 2014 20:32:33 +0000
- To: Roland Zink <roland@zinks.de>
- CC: "<ietf-http-wg@w3.org>" <ietf-http-wg@w3.org>
Hi Roland, thanks for the comments see in line On Nov 3, 2014, at 7:02 PM, Roland Zink <roland@zinks.de> wrote: > On 03.11.2014 14:26, Salvatore Loreto wrote: >> Users that select a proxy for optimization might still be required to accept the use of a proxy at their workplace or captive network of choice. > Do you suggest to create dynamic proxy chains? E.g. the proxy at the workplace go out to the optimization proxy to fulfill user requests. yes the idea is to investigate the possibility (i.e. the use case) to chain the proxies. The fact that a particular context forces the user to accept a particular proxy in between, should not preclude \ from using also proxies that the user has deliberately chosen to have in place for optimisation purposes or other kind of services (i.e. malware scanning). >> The trick here is balancing the requirements of different use cases. >> Proxies that provide privacy features need to be used exclusively, otherwise they could risk packet sniffing or pervasive monitoring by other nodes and risk the same sort of exposure that brought down the Silk Road. That's at odds with this scenario and we'd obviously need measures to ensure that this doesn't fail catastrophically. > For a workplace proxy it may not be an option to allow privacy features, so I guess there are conflicting requirements. That is an interesting point, that makes thinks even more tricky br Salvatore >> Any thoughts from the group would be great. >> >> br >> Salvatore >> >> [1] http://tools.ietf.org/html/draft-nottingham-web-proxy-desc-01 > Regards, > Roland >
Received on Monday, 3 November 2014 20:32:59 UTC