Re: [ppsp] tracker protocol review

Hi Yunfei,

Thank you very much for the comments.
About Encoding: 
Binary encoding of the protocol (both at the Request/Response layer or the Messaging layer) can be done without modifying the grammar of the protocol. EXI represents the Structure and the Content of an XML document as an EXI binary stream (header and body), therefore, it is just a matter of serialization of the XML Infoset into an octet stream, suitable for HTTP transfer encoding, enabling content negotiation to ensure that clients and servers understand EXI. Similarly, HTTP messages transporting EXI can  have a binary wire format, as being prepared for HTTP/2.0. We have already an implementation of the Tracker Protocol over SPDY, where the HTTP messages are binary encoded, with the advantages of multiplexing requests and header compression.

Cumprimentos/Regards,
Rui Cruz

Sent from my iPad2

On 05/11/2012, at 03:28, zhangyunfei <zhangyunfei@chinamobile.com> wrote:

> Hi Rui and Yingjie (Speaking individually) ,
>     I have read the updated tracker protocol draft and have the following comments:
> 1) Encoding issue: It's good to mention that PPSP-TP is intended to support binary encoding in section 4. However in section 5, 7 and 8, http+XML is still in the main page. Is this just for example? Or a general semantics desperation is enough? I think much of section 7.1 in current version can be placed in Appendix A. Are you intending to use EXI for the binary encoding and do you still need http for the underlying transport?
> 2)  Check the section 10 for the XML descriptions. Do we need the update of this part?
> 3)Security part, section 9.2, how does the content integrity check related to the tracker protocol should be better explained.
>  
> BR
> Yunfei
>  
> zhangyunfei
> _______________________________________________
> ppsp mailing list
> ppsp@ietf.org
> https://www.ietf.org/mailman/listinfo/ppsp