Re: Number of links towards a DPC (ITU)

Yeah, I got it wrong.
My idea was to use the SLS but still generate enough SLS for the linkset locally.
But this would cause the problem of mis-sequencing when links go down.
But that seems to be the only other alternative of using the entire 32 links.


-----Original Message-----
From: Brian F. G. Bidulock [mailto:[email protected]] 
Sent: 17 June 2013 17:02
To: Nemana, Satya
Cc: David Laight; Arvind Kumar; [email protected]
Subject: Re: [Sigtran] Number of links towards a DPC (ITU)

Satya,

Nemana, Satya wrote:                    (Mon, 17 Jun 2013 15:27:13)
> My 2 cents.
>
> The ITU spec may not say this, but why not do this.
>
> 1) Use the LSB to select the linkset within a combined linkset.

So, XXX0 goes to one linkset and XXX1 goes to the other.

> 2) Once the Linkset is selected, use the full SLS value to select one
>    of the links in the linkset.

On the first linkset, the entire SLS is XXX0.  There are only eight values here.  On the second linkset, the entire SLS is XXX1.  There are only eight values here.  So, 8 links per linkset, 16 links per combined linkset.

> In this way, we can still have 16 links in each linkset and combined 
> linkset can have 32 links and still loadshare across all the links.

Nope.  See above.

> Each linkset will still have 16 of the SLS values so, there won't be 
> any issues of un-even loadsharing even if some links go down in the 
> route to the final destination.

Each linkset has 8 values.

--
Brian F. G. Bidulock
[email protected]
http://www.openss7.org/