Hi Alex,

Thanks for sharing your definitions on LoWPAN terminology. I agree
with you that the current definition of "LoWPAN Mesh Node" is
incomplete. We must point out that it is a general term which includes
both a LoWPAN Host or a LoWPAN Router.

The link definition of RFC4861 is indeed useful. For the Use Cases
draft, we will follow the changes in the ND draft on what a "LoWPAN
Link" is.

It might be nit-picky, but didn't you mean "IP forwarding" when you
defined "IP routing"? And I disagree with your statement that "it has
two or more interfaces, each interface of the router has a different
IP link-local scope." I think that LoWPAN nodes mostly likely won't
have more than a single physical network interface...

Greetings,
Dominik


On Thu, Apr 23, 2009 at 6:14 PM, Alexandru Petrescu
<[email protected]> wrote:
> As I said previously, I have difficulty understanding some terms in the
> usecases draft.
>
> For example, it defines LoWPAN Host, LoWPAN Router and LoWPAN Node. However
> it doesn't say a LoWPAN Node may be a LoWPAN Host or a LoWPAN Router. (I
> think, commonly in IPv6, Nodes may be Host or Routers).
>
> It defines Mesh Under and Route Over; however I can't understand the
> following: does a "Route-Over" LoWPAN contain the same IP link-local scope
> or not?  If yes then this is very difficult to udnerstand, next to
> impossible.
>
> Also, another problem is that it defines the Route-Over and Mesh-Under in
> apple-to-orange terms ("boundaries" vs "radio range"), whereas Route-Over
> and Mesh-Under have often been easily compared, ie redapples-to-greenapples.
>
> Then on the use of 'single radio transmission' to define "Route Over" - what
> is single radio transmission more precisely?  One is tempted to believe all
> nodes reachable by a single radio transmission from a certain node are all
> within the same IP link-local scope... which is ok.
>
>>   Route Over
>>
>>      A LoWPAN configuration where the link-local scope is defined by
>>      those nodes reachable over a single radio transmission. [...]
>
> But because it doesn't say from _where_ are those nodes reachable (from
> another single certain node) then one can easily also think this:  B is
> within a single radio transmission of both C and A, yet it is in two
> different link-local scopes, which is of course not the intention.
>
>                   -----------------------
>                   |radiorange|radiorange|
>                   A          B          C
>
> Knowing the above, I'm trying to offer my own definitions of terms in the
> way I'd have written it.  Of course it may not fit all the views, but I
> thought I'd share:
>
> PHY repeating
>
>  The act performed by a device repeating a packet at PHY level.  It
>  doesn't inspect any field in any packet header.  It has one or several
>  interfaces, they are all in the same IP link-local scope.
>
> MAC forwarding
>
>  The act performed by a bridge - inspecting the dst MAC address of
>  incoming packet, exact-matching it to the first field in a table of
>  tuples [dst,nexthop]
>  MAC addresses and transmitting it to the identified nexthop.  In the
>  process, the src and dst MAC address of the packet are potentially
>  modified.  The device may have one or more interfaces but they are all
>  in the same IP link-local scope.
>
> IP routing
>
>  The act performed by a router - inspecting the dst
>  IP address of an incoming packet,
>  longest-prefix matching it to the first field of each entry
>  in a table of triolets
>  [dst,prefixlength,nexthop] IP addresses and transmitting it to the
>  identified nexthop.  In the process, neither the src nor the dst IP
>  addresses of the packet are ever modified.  It has two ore more
>  interfaces, each interface of the router has a different IP link-local
>  scope.
>
> IP link-local scope
>
>  The set of IP nodes reachable by an IP packet whose dst address starts
>  with 0xff02.  All nodes on a link are within the same IP link-local
>  scope, and nodes outside the link are outside the scope.  Transmitting
>  packets in the same link-local scope doesn't involve IP routing but
>  may involve MAC forwarding.
>
> Radio range of node
>
>  The radio range of a node is the set of nodes reachable from it, at
>  PHY layer.  For example, the radio range of a WiFi node is 50m in a
>  clear-sight atmospheric area, without physical obstacles and in good
>  weather.
>
>  Obviously, all nodes in same radio range of a node could be understood
>  to be in the same IP link-local scope.
>
>  However, a single-interface node can't be part of several IP
>  link-local scopes even if sits at the intersection of the radio ranges
>  of several other nodes.
>
> "Mesh-Under" LoWPAN
>
>  A LoWPAN configuration where the IP link-local scope is extended to
>  more nodes (beyond the radio range of a node) by means of MAC
>  forwarding exclusively.  All nodes in a "Mesh-Under" LoWPAN are within
>  the same IP link-local scope, despite the probable presence of several
>  PHY radio ranges - this alleviates the non-transitivity aspect.
>
> "Route-Over" LoWPAN
>
>  A LoWPAN configuration where the IP link-local scope is extended to
>  more nodes (beyond the radio range of a node) by means of transmitting
>  IP packets, using a novel IP routing method.  This novel method
>  runs on a LoWPAN Router (uses a single interface, instead of the
>  typical 2-interface router) and both interfaces are in the same IP
>  link-local scope (instead of the typical router having a different IP
>  link-local scope for each interface).
>
> "Routed" LoWPAN
>
>  A LoWPAN configuration where the IP link-local scope is not extended
>  beyond the radio range.  Certain nodes are manually designated as
>  special.  The radio range of each special node contains the set of
>  nodes in the same IP link-local scope, and each such range forms a
>  unique link.  Connecting two such links is performed by means of a
>  typical IP router (multiple interfaces, different link-local scope).
>
> Comments anyone?
>
> Alex
>
>
>
> _______________________________________________
> 6lowpan mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/6lowpan
>
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan

Reply via email to