request for review: multihome shim API

[Shinta Sugimoto <[e-mail]>]

Dear all,

I am co-authoring an internet draft [1] titled "Socket Application Program
Interface (API) for Multihoming Shim".  I appreciate if you could read the
draft and give us feedbacks.  Please note that we are planning to make this
document as an informational RFC, and we believe that getting feedbacks from
application programmers and reflecting those to the document are important.

[1] http://tools.ietf.org/html/draft-ietf-shim6-multihome-shim-api-08

Let me briefly explain what the draft is about.  As the title shows, the draft
defines a set of extensions to the socket API to provide applications
additional control on the locator management.  The API is for applications
that are running on top of an IP endhost that employs an ID/Locator separation
mechanism (we call it "shim").  Examples of shim protocols are HIP [2] and SHIM6 [3].
The most important feature of shim protocol is to provide stable identifier
to the upper layer protocols so that they can use it as an communication endpoint.
The shim allows applications to continue using the stable identifier even when
the host changes its attachment point to the network (e.g., in case of rehoming
event under multihome environment, or L3 handover under mobile environment).
The implication of the shim protocol running inside the IP layer is that it hides
locators from the upper layer protocols.  That is certainly a good feature for
achieving session continuity as mentioned above.  However, we believe it may still
be useful for applications to acquire control on locator management.  Locator
management includes specifying locator for outbound IP traffic and knowing the
locators that appear in the wire packet formats.  Without the API, applications
cannot perform locator management.  One example where application may benefit
by having locator management is that application may want select particular
path (e.g., for achieving better transport performance) under multihomed
environment.

Any feedbacks are very welcome.  Thank you for your attention.

[2] http://www.ietf.org/rfc/rfc5201.txt
[3] http://tools.ietf.org/html/draft-ietf-shim6-proto


Regards,
Shinta