Shell Standard and Autoload Need to Support Breakout Cables

[alonagetzler]

@chris.g:
Breakout cables are very common. For example, you may use a QSFP+ to turn a single 40gig interface into four 10 gig interfaces using a breakout cable.

The old autoload driver uses the outputs of both the “ show chassis hardware “ to get the Interfaces subresources. This is the output of the chassis hardware for the QSFP+ on an MX:

FPC 8 REV 12 750-057064 ACAM7152 FPC3-SFF-PTX-1X
CPU BUILTIN BUILTIN SMPC PMB
PIC 0 REV 24 750-052081 ACAM8555 96x10GE/24x40GE/8x100GE QSFP+ PIC
Xcvr 0 REV 01 740-054050 INEBU1151641 QSFP+-4X10G-LR
Xcvr 1 REV 01 740-054050 INFAH0491147 QSFP+-4X10G-LR
It uses this output to create the Interface subresources so it would create et-8/0/0 and et-8/0/1. But with the QFFP+ and breakout cables you would have another level of interfaces under the physical port/optic on the router and this can be seen via the show interface terse command.

Here’s the show interface terse output for that same FPC/PIC/Optics pairings.

et-8/0/0:0 up down
et-8/0/0:1 up down
et-8/0/0:2 up down
et-8/0/0:3 up down
et-8/0/1:0 up down
et-8/0/1:1 up down
et-8/0/1:2 up down
et-8/0/1:3 up down
So as you see, I need to add the above list of interfaces into Cloudshell instead of just the 2 interfaces that would be created by using the show chassis Hardware command.

This would need to be tweaked since the logic is now SNMP, but you get the point. Although this was shown for Juniper, this is pretty common.

raised by Chris G in the idea box:
http://community.quali.com/content/idea/586/shell-standard-and-autoload-need-to-support-breako.html