Re: [6tisch] [6lo] Compression Reference and Coalescence

[Tengfei Chang <[e-mail]>]

Yes, Pascal. What you are going to say is correct!

I agree +1

Tengfei

On Tue, Jan 26, 2016 at 6:06 PM, Pascal Thubert (pthubert) <
[e-mail]> wrote:

> I see, and yes, that is misleading.
>
>
>
> I need to change that to say the reference to the first is the source of
> the packet and then the reference for each entry is the previous one when
> uncompressed (not the first as you suggest but, the one just before it).
>
> So you uncompress an address and it becomes reference to the next that you
> uncompress, etc… I will raise an ticket. . You’ll find that the algorithm
> in “5.4.  Popping Headers” decodes exactly that (well hopefully).
>
>
>
> Do we agree?
>
>
>
> Pascal
>
>
>
> *From:* Tengfei Chang [mailto:[e-mail]]
> *Sent:* mardi 26 janvier 2016 17:56
> *To:* Pascal Thubert (pthubert) <[e-mail]>
> *Cc:* [e-mail]; [e-mail]
> *Subject:* Re: [6lo] Compression Reference and Coalescence
>
>
>
> Nice to hear that's what you intended to do!
>
>
>
> In  the beginning of section 4.3:
>
> ...
>
>
>
>    Once the address of the source of the packet is determined, it
>
>    becomes the reference for the compression of the addresses that are
>
>    located in compressed RH3 headers that are present inside the IP-in-
>
>    IP encapsulation in the uncompressed form.
>
> ...
>
>
> When I am understanding the sentence, I am feeling like all the
> compression of addresses in RH3 will be compressed according to source
> address in IPinIP, which only the first 8 bytes can be elided.
>
>
>
> If agree with the reasonable one, it should be the first address in first
> RH3 is compressed according to the source address in IPinIP.
>
> Then, all rest address in RH3 will be compressed according to the first
> address in RH3.
>
>
>
> Make sense?
>
> Tengfei
>
>
>
> On Tue, Jan 26, 2016 at 5:23 PM, Pascal Thubert (pthubert) <
> [e-mail]> wrote:
>
> Hello Tengfei:
>
>
>
> The draft is meant to express exactly what your reasonable case shows.
>
> There is an example of that in appendix.
>
> What exactly did I write improperly that it can be understood otherwise?
>
>
>
> Take care ;
>
>
>
> Pascal
>
>
>
> *From:* 6lo [mailto:[e-mail]] *On Behalf Of *Tengfei Chang
> *Sent:* lundi 25 janvier 2016 12:00
> *To:* [e-mail]; [e-mail]
> *Subject:* [6lo] Compression Reference and Coalescence
>
>
>
> Dear all,
>
>
>
> I have some concern on the compression Reference and Coalescence.
>
>
>
> In the 6lorh draft, all the hops in RH3 are compressed according to the
> compression reference (the source address, root usually). And when doing
> Coalescence, the process is kind of like taking the first address in first
> RH3 as the reference. The inconsistency between compression reference and
> Coalescense may waste some bytes in some cases.
>
>
>
> For example:
>
> A packet is issued by root with a compressed RH3 along an A->B->C->D
> source route path.
>
>
>
> The nodes address are:
>
> root:      bbbb::0000:0000:0000:0001
>
> node A: bbbb::1111:2222:3333:0001
>
> node B: bbbb::1111:2222:3333:0002
>
> node C: bbbb::1111:2222:3333:0003
>
> node D: bbbb::1111:2222:3333:0004
>
>
>
> According to the 6lorh draft: all hops in RH3 will be compressed according
> to reference, which is the root. The  Packet received by node A is:
> ----------------------------
>
> Type 3 RH3-6LoRH Size = 2  1111 2222 3333 0001
>                                            1111 2222 3333 0002
>                                            1111 2222 3333 0003
>
>
>
> And which maybe more reasonable packet would be like :
>
> ----------------------------
>
> Type 3 RH3-6LoRH Size = 0  1111 2222 3333 0001
> Type 0 RH3-6LoRH Size = 1  02
>                                             03
>
> Which means the first hop in first RH3 entry is compressed according to
> the reference(which is the root in this case) and the rest hops are
> compressed according to the first hop in first RH3 entry. For me, this
> compression way is more consistent  with the way when doing coalescence.
>
>
>
> What do you think?
>
> Tengfei
>
>
>