CC2430: CC2430 ZStack-1.4.3-1.2.1 Cskip and max Routers in tree topology

Hi Joaquim,

Joaquim Martinez said:

since children nodes can be Routers and children Routers

Can you please explain your reasoning behind this?  My understanding is that children (or associated devices) are either routers or end devices.  MAX_ROUTERS is reserved for router association, and MAX_CHILDREN-MAX_ROUTERS is reserved for end device association.  Thus MAX_ROUTERS defines how many routers you can have "horizontally" at any given depth for a node.  Note that the Developer's Guide includes recommendations on variables which must be considered if altering the default values.  There are also practical limitations to implementing large networks regardless of theoretical maximum node capabilities.

Regards,
Ryan

Hi Ryan,

Exactly the same that you say: "My understanding is that children (or associated devices) are either routers or end devices".

Now I understand that MAX_ROUTERS is "how many routers you can have "horizontally" at any given depth for a node". In addtion if I understand well, MAX_CHILDREN-MAX_ROUTER are the End Devices for each Router, isn't it? 

My concern is, for example if Lm = 5, Rm = 2 and Cm = 3. So, let me expose the next points:

1. The coordinator can have associated 1 ED and 2 Routers.

2. Each Router can have 1 ED and 2 Routers.

3. Only a ED can be addressed for Coor and Routers.

4. According to the attached diagram, in the tree topology with this configuration I can have 1 Coor, 30 Routers and 31 ED, it is correct? I have calculated these number according to the Cskip values and points 1, 2 and 3.

5. But according to the Cskip equation, up to 95 nodes (1 coord + 94 nodes) can be addressed, whereas according my diagram I have 62 nodes (1 coor + 30 routers + 31 end devices).

What is wrong?

Thanks

Joaquim

I think your diagram is missing one depth, as D5 should also have routers.  This results in a count of 46 routers (2 routers at depth 1, 4 routers at depth 2, 8 routers at depth 3, 16 routers at depth 4, 32 routers at depth 5), 1 coordinator, and 47 end devices (1 for each router and the coordinator) for a total of 94 which is much closer to the Cskip equation.

Regards,
Ryan

Thanks for the answer. 

I though that at last Cskip level (d5) only can have ED, because the Cskip vakue is 0, so no address can be assigned and thus is not possible to have Routers. Please, see the attached image extracted from a Zigbee book. I highlighted in blue/red the explanation.

Concerning to the number of Routers and ED, sorry but I don't understand yet: adding (2 routers at depth 1, 4 routers at depth 2, 8 routers at depth 3, 16 routers at depth 4, 32 routers at depth 5) the result is 62 Routers (2+4+8+16+32 = 62). In addition, if each Router and Coordinator have 1 ED, there are 63 ED... so the total nodes are: 1 coor + 62 routers + 63 ED = 126 nodes -> far away from the 94 nodes of the Cskip equation. 

So I have still these two concerns. 

Thanks.

Joaquim.

My apologies for the poor math!  And you are correct about the topology.  Please allow me to retry: with an excerpt from Zigbee Wireless Networking by Drew Gislason:

Using the formula maxNodes = 1 + Rm x Cskip(0) + (Cm-Rm) = 1 + 2 x 46 + 1 = 94, which is the same as what you've calculated before.

I believe I found the flaw in your diagram.  Do you see how in d4 you have two end devices connected per router, but no routers?  I think each router in d4 should branch off the same as d0/d1/d2/d3 with two routers (child-less this time) and one end device. This would make for 46 individual nodes on the left side of the tree, thus doubling for the right side and adding the coordinator and its own end device would result in a total of 46 x 2 + 2 = 94 nodes, the same as what is calculated.

I hope this helps,
Ryan

No worry.

But in d4, I have already 2 Routers and 1 ED by Router: in total 8R and 8 ED. Do you mean that d5 depth must have 8R and 8ED? For me, is the only way to have 46 nodes by branch (in total 46 * 2 + 2), but this go against to have 2 child Routers by Router parent and 1 ED all them in "Horitzontally" way, isn't? If in d5 depth I put 2 Router (from d4) and 1 ED, then we have 24 nodes at d5, and in total 110 nodes.

Sorry, but I am going crazy...

Joaquim.

So, we can assume that:

1. Not all Routers can have 1 ED. If yes, the key here are the ED, because the Routers can have 2 child Routers, isn't?

2. It is not necessary or is not possible to have the same amount of EDin each branch of the tree. I mean that for instance, in D1 only a branch of the tree will/can have one ED, the other branch do not, or in D4 only 8 Routers (of 16) will/can have 1 ED each.

3. If above is true, that means the tree can be "asymmetric", I expected it has to be symmetric. 

4. Finally, I understand that Cskip is a maximum value, and the addresses define the topology of the branches, either asymmetrical or symmetrical, isn't?

Hi Ryan,

Now all is clear. The ED for each Router or Coordinator has to be on the next depth of their parent... In my diagram the ED were in the same depth... This is the error... 

Thank you very much for your support.

Joaquim.