On 5/16/2011 9:32 PM, Hao Zhou wrote:

...

>> Sam Hartman wrote:
>>> I'd like to confirm that code is in use both by implementations of
>>> eap-fast v1 and v2.
>>
> [HZ] Yes. Type Code 43 is being used by EAP-FAST v1 and widely deployed.
>  
>>   As a backup question: Are there *any* implementations of v2?
>>
> [HZ] Not right now. Once this draft is finalized, it will be.
>  
>>   The draft does not make it clear if this is the case.  Can the authors
>> step in and give their opinion?
>>
>>> Does the current text mandate support for eap-fast v1 as well as v2?
>>
> [HZ] The draft doesn't mandate support for v1 and v2, it's up to each
> implementation. How, ever, due to the small changes from V2 to v1, I suspect
> most of them could support both easily. Doesn't running code mean something
> in IETF?

Let me get this straight: are you claiming that changing the EAP type
will break the "running code" but making the same "small changes" and
keeping the same EAP type won't?

>  
>>   Yes and no.  Section 3.1 says:
>>
>>    The version negotiation procedure guarantees that the EAP-FAST peer
>>    and server will agree to the latest version supported by both
>>    parties.  If version negotiation fails, then use of EAP-FAST will not
>>    be possible, and another mutually acceptable EAP method will need to
>>    be negotiated if authentication is to proceed.
>>
>>   This makes it *possible* for an implementation to support v2 only.
>> This will require starting version negotiation for EAP-FASTv2, and then
>> switching to a different EAP method.
>>
>>   Implementations traditionally have found it difficult to start one EAP
>> method, and then to switch to another one.  This means that v2-only
>> implementations may be difficult to deploy in practice.
>>
> [HZ] Agree. Considering all devices out there supporting EAP-FASTv1 and it
> may take a while for them to support the new method, most server
> implementations would probably support both.

New devices are created every day.  I don't think that it's too much of
a stretch to think that device vendors would prefer to support only one
tunneled method; after all, isn't that one of the purposes of this
exercise?

>>> Is it expected that most implementations will support v1 and v2?
>>>
> [HZ] See above.
> 
>>> Is it desired that people be able to create a v2 only implementation?
>>
>>   I will partially avoid those two questions, and say that it should be
>> possible to deploy only the EMU tunneled method.
> [HZ] Realistically, until the new tunnel method is published and adopted by
> all servers and supplicants, implementation would have to support older
> tunnel methods, including EAP-FASTv1. By retaining the EAP-FAST type,
> customers wouldn't have to be educated with another new EAP type and
> validate its security, and they would have a smoother migration path, from
> v1 to v2. 

Is the claim here that customers are so stupid that they would deploy a
new version of EAP-FAST without validating its security?

> Implementers could reuse the same code. 

This is a non-issue. The same code can be reused regardless of the EAP type.

> I thought one of the
> reasons EAP-FASTv2 is adopted is because of its existing code and
> deployment, with small changes to meet EMU requirements. 

Again, I'm confused: you wrote above that there are no implementations
in existence.  If we adopted EAP-FASTv2 "because of its existing code
and deployment" then we must certainly be deluded since there is no code
and there are no deployments.

> Changing the method
> name and type would totally negativate that.

I suspect that all it would really negate is the marketing opportunity.
 OTOH, the Cisco marketing machine is truly awesome while the IETF's,
not so much ;-).  Would vendors (other than the employers of the
authors) be more likely to implement EAP-FASTv2 that the IETF Tunneled
EAP Method?  Would customers be more likely to deploy it?  I don't know,
but I think that it's worth considering.  AFAICT the difference between
the two options in terms of development and operational costs is very
close to zero, though.

...

<<attachment: gwz.vcf>>

_______________________________________________
Emu mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/emu

Reply via email to