[Jool-list] Jool use with IPv6 island connected to IPv4 network

[Alberto Leiva]

> I understand the mapping concept, and this question was for when
> G64 already created the mapping for UDP between the IPv6 address
> and the IPv4 one (when using a pool4).

Are you sure you are not mistaking BIB and pool4?

pool4 is just the transport addresses the NAT64 can use to mask
IPv6 nodes. What's important is that it doesn't define *which* IPv6
nodes are masked. That is BIB's responsibility.

pool4 is just the pool of available v4 masks. The masks themselves
are BIB entries.

To create a UDP mapping/mask/BIB entry you have to do something
like

    jool --bib --add --udp fg01::3e52:10b1:6fc8:f0f1:94f8#80
192.168.0.100#80

> when Jool creates a mapping between one IPv6 and an IPv4 (this
> would only happen on A6-initiated traffic), I would like for it
> to ask to the DHCP for an address instead of randomly grabbing
> one from the pool4.

By core nature, NAT64 does not reserve addresses; it reserves
*transport* addresses. In the command I wrote above, only port 80
of address 192.168.0.100 is reserved. Other ports of 192.168.0.100
can potentially belong to other IPv6 nodes.

If you need to reserve addresses, you definitely need SIIT.

> Yes, I got that. But then, when we define a pool4, it stops doing
> this when it maps an IPv6 traffic to an IPv4. Or maybe I don't
> understand the use of pool4?

pool4 does not map any v6 traffic to v4; see above.

> Agreed, but I think the problem comes from G64 that does not know
> that indeed .54 traffic is expected to be for it (more
> specifically for one of its services).

I don't think so; from the way I see it, it's a layer 2 thing.

If this is C4's routing table:

    $ ip route
    default via 192.168.0.100 dev eth0  proto static
    192.168.0.0/24 dev eth0  (...)  src 192.168.0.104  (...)

Then it doesn't have any way to know that it's supposed to send .54
packets via 192.168.0.100. The link scoped route (192.168.0.0/24)
is more a more specific match to 192.168.0.54 than the global one
(0.0.0.0/0) so it attempts to find .54 as a neighbor (because the
192.168.0 route is link-scoped).

Assuming that .54 is not in its neighbor table ($ ip neigh), it
sends an ARP request: "Who is 192.168.0.54?," and since G64 does
not have the address in its interface, nobody answers. Therefore,
C4 thinks .54 does not exist. Therefore, it never puts the packet
on the network. G64 cannot receive it in the first place.

If you are seeing the packet in G64 despite the above problem is
because the earlier ping managed to hack a bogus neighbor table
entry intr C4, mapping G64's MAC address (sort of incorrectly) to
.54. Assuming you can live with this, G64 will receive the packet
and Jool will have no problem snatching and translating it. That's
the reason why adding the 192.168.0.54/32 route to C4 works; it
forces C4 to skip the neighbor discovery step. Because it's not
link scoped and is more specific than 192.168.0.0/24.

But this is all assuming you already added a BIB entry for whatever
port C4 wants to talk to! If that mapping does not exist, even if
Jool catches the packet, it will be unable to tell where it is
headed to.


On Fri, Jul 29, 2016 at 9:10 AM, Romain Bazile <romain.bazile at ubiant.com>
wrote:

> Hi,
>
>
> First of all, thanks for the answers!
>
> Regarding your comments, it looks I did not made myself very clear! I
> answered below.
>
> This property of remembering mappings is what makes NAT64
> "stateful". Its advantage is that you're using a few G64 addresses
> to grant a potentially large and dynamic number of IPv6 nodes
> access to IPv4 ones. The drawback is that IPv4 nodes cannot start
> communication with v6 ones because G64 doesn't initially know who
> they are talking to.
>
> (Unless you predefine mappings yourself. You can find more
> information at https://jool.mx/en/bib.html and
> https://jool.mx/en/usr-flags-bib.html)
>
> I understand the mapping concept, and this question was for when G64
> already created the mapping for UDP between the IPv6 address and the IPv4
> one (when using a pool4).
> This drawback is exactly why I would want some kind DHCPing to happen:
> I.E. when Jool creates a mapping between one IPv6 and an IPv4 (this would
> only happen on A6-initiated traffic), I would like for it to ask to the
> DHCP for an address instead of randomly grabbing one from the pool4. And
> this mapping would have the same TTL as the allocation from the DHCP server.
>
> I read the doc for the bib entries, it would be perfect, except my only
> problem with those is that I would have to manually add every device as
> soon as it appears on the network. And I'm not sure how to do that (yet!
> Maybe I could broadcast a ping or use neighbour discovery to find IPv6
> devices and attribute them a fixed entry, all from within a script that
> would also push new bib commands).
>
>
> One of the downsides of SIIT is that this "predetermination" of
> address mappings is not going to be friendly with your "Every IPv6
> device has a somewhat unpredictable IP where the prefix is
> fg01::/64" requirement.
>
> You can either find a way to impose more order to your address
> assignment policies (so they will be friendlier to SIIT's
> constraints)
>
> Yes, this is why I can't use SIIT. I can't impose more constraints on the
> IP of these devices...
>
>
> or you can try Taiga (http://www.litech.org/tayga/).
> Tayga is a SIIT/NAT64 hybrid in that it can assign v4 addresses to
> IPv6 nodes on the fly, and then the entire address (not just an
> address and port) are assigned to the IPv6 node. I can't vouch for
> Tayga because I don't use it and hasn't received updates in a long
> time, but perhaps you can squeeze something out of it.
>
> I tried it! But it doesn't compile out of the box for kernel 4.1.15. I
> didn't have time to investigate where the problem comes from.
>
>
>
> Alternatively, you might want to open this hybrid operation mode as
> a feature request for Jool: https://github.com/NICMx/Jool/issues
>
> That's an idea! I'll try and define more precisely the expected behavior
> before putting it up in Github.
>
>
> > Here, I reckon that the gateway does not know that packets on the
> > network with the 192.168.0.54 address are in fact destined to
> > her, because when I manually add this IP to eth0 (with ip addr
> > add 192.168.0.54/24 dev eth0 ), suddenly, it works!
>
> A NAT64 is a device that masks IPv6 nodes behind it using its own
> IPv4 address (or addresses).
> An SIIT is a device that renames your IPv6 nodes, each using a
> separate IPv4 address, without them realizing it.
>
> Pay special attention to the part where I say "using its own IPv4
> address". NAT64 does not rename v6 nodes like SIIT does; it
> pretends it itself is the IPv6 nodes it is masking. That's why they
> are called "masks" :)
>
> So yes; the v4 nodes need to be fooled into thinking that G64 *is*
> the v6 network. One way to do that is by adding the addresses to
> G64's interface. (So G64 answers ARP "neighbor" requests to these
> packets.)
>
> Yes, I got that. But then, when we define a pool4, it stops doing this
> when it maps an IPv6 traffic to an IPv4. Or maybe I don't understand the
> use of pool4?
>
>
> > How can I tell my eth0 interface that its should be concerned by
> packets adressed on the 192.168.0.54 (which it used to send packets
> beforehand since pinging worked), but without statically
> configuring it with this IP address?
>
> You need to realize that 192.168.0.54 belongs to C4's network
> (192.168.0.54/100), so C4 is expecting .54 to be one of its
> neighbors. For this reason, it doesn't intend to send the packet
> via any gateways.
>
> Agreed, but I think the problem comes from G64 that does not know that
> indeed .54 traffic is expected to be for it (more specifically for one of
> its services).
>
>
> That said, you can always instead inform C4 that any traffic
> towards 192.168.0.54 should be routed via G64:
>
>     romain at c4$ ip route add 192.168.0.54/32 via 192.168.0.100
>
> But perhaps by this point you have realized that you should more
> likely redesign the setup :p
>
> Yes, and this is something I would like to avoid.
>
>
> > Can jool automatically and dynamically requests a new IPv4 from a
> DHCP server for an IPv6 device hidden behind?
>
> No, but why would you want to do that? (maybe I'm missing
> something)
>
> I explained in the beginning of this email why I want this.
>
>
> > Can jool automatically set up new address on the eth0 interface
> (when they are allocated by the DHCP server)?
>
> No, but can't a shell script do it?
>
> It definitely could! I think Michael's proposal is quite interesting, I'll
> start playing with macvlan to see how it behaves with the dhcp and
> everything.
>
>
>
> By the way: It seems that you're having trouble understanding these
> concepts from the documentation. Any tips on how to improve it?
>
> I don't yet. I think that when I have better grasp on all those things I
> will be able to find areas that can be improved (if there are any).
>
>
>
>
> Romain Bazile
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail-lists.nic.mx/pipermail/jool-list/attachments/20160729/1a829ee4/attachment-0001.html>