Re: [IPsec] [I2nsf] Review of draft-ietf-i2nsf-sdn-ipsec-flow-protection-03 (Section 1)

[Paul Wouters <[e-mail]>]

On Tue, 27 Nov 2018, Gabriel Lopez wrote:

> Hi Paul, all
> 
> Please find attached some answers to your comments. Let’s go section by section, it will be easier to follow the discussion.
> 
>
>       El 18 nov 2018, a las 7:52, Paul Wouters <[e-mail]> escribió:
> 
> 
> General comments:
> I'd like to see "Case 1" and "Case 2" replaced with more descriptive terms. I keep losing track of which case is which.
> Perhaps call them IKE Case and IKE-less Case ?
> 
> 
> [Authors] We agree. What about "IKE-full" and "IKE-less" cases?

Hmm, I find "IKE-full" a bit odd (also because in English you would
probably say "full-IKE" and not "IKE-full". Maybe another word could
be "IKE-managed" or "IKE-controled" or "IKE-delegated" ?

> [Authors]  Thanks for the suggestion. Certainly the text is not reflecting the existence of those techniques alleviating the task of establishing IPsec SAs. Nonetheless,
> we have to clarify that this draft is not only guided by the SDWAN scenario and, in fact, it is intended to be applicable in other foreseeable SDN deployments.  With
> this in mind, we propose the following changes in the text:
> 
> 
> s/"Typically, traditional IPsec VPN concentrators and, in general,
>    entities (i.e. hosts or security gateways) supporting IKE/IPsec, must
>    be configured directly by the administrator.  This makes the IPsec
>    security association (SA) management difficult and generates a lack
>    of flexibility, specially if the number of security policies and SAs
>    to handle is high.  With the growth of SDN-based scenarios where
>    network resources are deployed in an autonomous manner, a mechanism
>    to manage IPsec SAs according to the SDN architecture becomes more
>    relevant.  Thus, the SDN-based service described in this document
>    will autonomously deal with IPsec SAs management." /
>    
>    "With the growth of SDN-based scenarios where
>    network resources are deployed in an autonomous manner, a mechanism
>    to manage IPsec SAs according to the SDN architecture becomes more
>    relevant."
>    
> s/"An example of usage can be the notion of Software Defined WAN (SD-
>    WAN), SDN extension providing a software abstraction to create secure
>    network overlays over traditional WAN and branch networks.  SD-WAN is
>    based on IPsec as underlying security protocol and aims to provide
>    flexible, automated, fast deployment and on-demand security network
>    services." /
>    
>    "An example of usage can be the notion of Software Defined WAN (SD-
>    WAN), SDN extension providing a software abstraction to create secure
>    network overlays over traditional WAN and branch networks.  SD-WAN is
>    based on IPsec as underlying security protocol and aims to provide
>    flexible, automated, fast deployment and on-demand security network
>    services. Other examples are host-to-host or full-mesh encryption. 
>    Thus, the SDN-based service described in this document will autonomously deal with IPsec 
>    SAs management.

works for me.

>         The analysis of the host-to-gateway (roadwarrior) scenario is TBD.
>
>       Does this sentence mean this is to be done in this document or is it out
>       of scope for this document (but it does allow it to be added later) ?
> 
> [Authors] We do not see a clear use case for roadwarrior scenarios based on SDN, the proposal
>   is to remove it from this document. 

The case I can think of is if this is a "site to site" connection but
one site is on dynamic IP (eg cable modem or 4g/5g connection) ?

The only difference in yang would be to instead of a static ip or
hostname, to also allow a value of "any IP". In libreswan we use
the string "%any" (and %any4 or %any6) but one could also use 0.0.0.0
to mean that. I don't think it needs any other special handling (other
then making sure an endpoint 0.0.0.0 should not use an IP based
identity.

Paul