Re: [v6ops] new draft: draft-cui-v6ops-lte-lw4over6

Hi Mikael,

Thanks for your comments! Please see inline.
Please don't hesitate to point out if there is any mistake. Thank you!

//sorry for send it again (used an wrong mailbox previously)

On 2014-2-13, at 下午9:58, Mikael Abrahamsson wrote:

> On Thu, 13 Feb 2014, fred@cisco.com wrote:
> 
>> A new draft has been posted, at http://tools.ietf.org/html/draft-cui-v6ops-lte-lw4over6. Please take a look at it and comment.
> 
> This document makes me very confused.
> 
> Traditionally, the eNodeB has basically done GTP-RLC (or whatever protocol is used on the radio layer) conversion. All the setup has been done using 3GPP protocols.

[Qi] But the eNodeB still needs an IP address, which is allocated through DHCP. 

> 
> This document proposes to use DHCPv6oDHCPv4 with the eNodeB as a *client*.

[Qi] When the eNodeB launches, it needs to setup the default bearer to get configuration. In this process, DHCPv6 is used for IPv6 configuration. So we are thinking if DHCPv4oDHCPv6 could be used to configure the eNodeB with an IPv4 address over the IPv6 network.

> When I read the document, I don't understand what communication is done within the GTP tunnel,

[Qi] Typically, it's the IPv4-GTP-IPv6 encapsulation and just data forwarding within the GTP tunnel. 

> what is done outside the tunnel,

[Qi] We think it could be the private-public IPv4 address translation (including source port mapping on the eNodeB), and the mapping lookup on the PGW when there are packets back to the UE from the IPv4 Internet.

> and what role the eNodeB has in this.

[Qi] The eNodeB could play the role of the "gateway" for the UEs, i.e. allocating private IPv4 addresses to the UEs and do the private-public IPv4 translation within a restricted port range. 

> 
> In 4 for instance:
> 
> "The architecture described here addresses a typical use case, where
>  an eNodeB's uplink supports IPv6 only and a UE using IPv4 in this
>  eNodeB wants to access IPv4 Internet.  The network architecture is
>  shown in Figure 1.  In this scenario, the UE can only use the IPv6
>  network to access IPv4 services, so IPv4 services must be configured
>  over IPv6."
> 
> I don't understand this statement. It's completely possible to set up an GTP tunnel over IPv6, which then can carry both IPv4 and IPv6 traffic. So it's not problem to support UEs with IPv4, IPv6 or DS uplinks, in current 3GPP architecture, over a pure IPv6 mobile backhaul network.
> 
> I'm very confused what problem this document tries to solve.

[Qi] The proposal is trying to use tunneling mechanism (lw4over6) to help LTE networks and users to transition to pure IPv6. We intend to make use of the GTP tunnel for encapsulation and avoid modifications on the UEs. In addition, the PGW can avoid large scale NAT at the cost of assign shared public IPv4 addresses to the eNodeBs. 
The difference to the usage of xlat464 in LTE would be: no changes to UEs, and no large scale NAT (translation) on the PGW.

Maybe we are not depict the purpose precisely. Please feel free to comment. Thanks again.

Best Regards,
Qi

> 
> -- 
> Mikael Abrahamsson    email: swmike@swm.pp.se
> 
> _______________________________________________
> v6ops mailing list
> v6ops@ietf.org
> https://www.ietf.org/mailman/listinfo/v6ops

Dear ,

Best Regards,
Qi Sun