- From: James P. Salsman <bovik@best.com>
- Date: Sat, 16 Sep 2000 17:44:56 -0700 (PDT)
- To: paf@cisco.com
- Cc: discuss@apps.ietf.org, ietf-mmms@imc.org, ietf@ietf.org
Patrik, Thanks for your message: >... I want a more specific list of documents that you are to create, Two documents would probably make sense: (1) End-to-End Internet Services for Mobile Devices Scope: Specifications and interoperability guidelines for end-to-end mobile IP connection and transport services required for support of standard Internet messaging (e.g., TCP, UDP, ICMP, with general descriptions of such wireless Internet services that already exist), along with guidelines for TCP operation during indefinite wireless link downtime, determining bandwidth and link conditions, and a very few hardware-level specifications (e.g., serial cable PPP access for connected devices.) (2) Internet Multimedia Messaging Services for Mobile Devices Scope: Specifications and interoperability guidelines for multimedia messaging using MIME-based Internet messages and related services on mobile devices, including application and session services (e.g., IMAP and POP with guidelines for their use in a wireless context, SMTP to and from wireless devices), and specific guidelines for MIME message content types and encodings for interoperable, unencumbered media formats (such as the open formats in the RTP Profile RFC 1890.) > and the list of issues you are to discuss should also say what > is _NOT_ in scope of the wg. Not in scope: No new protocols or formats will be created. No specifications with any intellectual property encumbrances will be incorporated. No content negotiation or preference-profile services will be defined (as those are already within the scope of the conneg and rescap working groups, and the W3C Content Capability and Preference Profile (cc/pp) working group), although of course those groups and their documents would all be cited in the Internet Multimedia Messaging Services for Mobile Devices document. > You should by the way define a "mobile client"... Good point. > I felt the discussion at the BOF ended up with "a client which > have very limited bandwidth to the network, often connected to > many different internet providers, limited memory and screen" The bandwidth limitations of wireless devices these days are often not too bad. Ricochet/Metricom is usually never more than twice as slow as a typical telephone modem. (By the way, at least one person thinks I'm affiliated with Ricochet/Metricom; I am not.) Some of the experimental 2.5 GHz IEEE 802.11 protocol variants work with cells many miles in radius and are capable of shared bandwidths many times greater than a T1 -- for example: http://www.overlan.com/pages/prod/rf11plus.html (I'm not affiliated with C-Spec/Overlan, either.) The real bandwidth issue isn't the typical bit rate; instead, wireless devices will for many reasons have occasional sudden bandwidth dips or outright link failure for extended periods of time. TCP parameters need to be configured to be more forgiving of those situations; that would be addressed in detail in the End-to-End Internet Services for Mobile Devices document. Memory limitations, while real, are still keeping pace with Moore's Law in general, and should be expected to continue for some time. It might make more sense to say that mobile devices are less likely to have any kind of mass storage, and rely upon existing email session mechanisms for rejecting messages of too great a length, or with components in unreadable formats. Those concerns would be well within the scope of the Internet Multimedia Messaging Services for Mobile Devices document. Screen-size is a valid user interface concern, being addressed by the rescap, conneg, and W3C CC/PP working groups, and I would be very reluctant to suggest that topic for any MMMS specifications. Cheers, James
Received on Saturday, 16 September 2000 20:46:03 UTC