I don't disagree.
We need to make recommendations along your thoughts and let SDOs and operators
decide how to implement their networks.
By the way, a single-sign-on network is also a walled garden right? The walled
garden is between the parties that aggregate around the identity service
provider. I am thinking Passport (especially), I am thinking Liberity
Alliance, I am thinking Open-ID.
In that vain it is also worthwhile to note that an operator may choose to
bootstrap secruity associations from EMSK between a MN accessing its network
and third pary Application Service Providers who have a relationship with the
Operator. In such a relationship the MN does not have to reauthenticate with
the Application Service Providers. This is an example of a single sign on.
The only way to elliminate any walled gardens is to have the mobile have its
own relationship with each application provider. This has advanatages and also
Sent: Tuesday, March 25, 2008 3:50 AM
To: Avi Lior; aboba(_at_)internaut(_dot_)com; ietf(_at_)ietf(_dot_)org
Subject: RE: IETF Last Call on Walled Garden Standard for the Internet
Avi Lior wrote:
Here I agree with you fully: this is an extremely bad idea.
Architecturally linking application security to the link layer is
just bad engineering, and hinders the ability of link layers and
applications evolve independently of each other.
Lets start with this: Any application?
Well, at least applications which are not inherently (*) tied
to a specific access network.
In other words: if it simply doesn't make any sense to use
the "application" from a different link or access network,
then tying it to the link layer authentication might be one
Otherwise, it's a bad idea.
(*) Inherently: by their nature -- and not e.g. just by
current business structures (which are likely to change due
to mergers, acquisitions, divestiture, etc.) or SDO
boundaries (both users, access providers, and service
providers are, over the time, likely to be interested in
network access technologies from multiple SDOs).
The emsk-hierarchy document should not give higher layer
applications as an example use case; instead, it should
this is a bad idea, and recommend that keys derived from
authentication should be used solely for "link-layerish" things
(such as link layer handoffs; Mobile IP is a borderline
Mobile IP is an application. So I guess you are okay with some
Someone mentioned DHCP as one "application" which is
inherently tied to a specific access network/link.
If you want to use Mobile IP to provide mobility only within
a single access network -- and assume that neither you nor
your customers will ever be interested in other access
technologies in the future (or that mobility to e.g., IETF
WLAN is either unimportant, or handled by some other
mechanisms), then you could tie Mobile IP and link layer
authentication. Otherwise, I'd recommend making it access independent.
IETF mailing list