Hi, Joe

Thanks for the follow-up and further review, much appreciated. Please also see 
my reply inline.

[JMC] This flexibility does make sense.  That said, now it seems that the 
augmentations in this draft could apply both to the SDN controller and the PEP, 
correct?
[Qiufang] Yes, that is exactly what have been proposed in the current version 
of the draft.
I get the former as part of the new Step 1.  But the latter wasn't clear to me 
in Step 5.  Ultimately, I think an example of both cases (MAC/IP and group) 
could be helpful along with some additional text to indicate what type of ACL 
config may exist on the PEP.
[Qiufang] Good suggestion. Will expand step 5 and add an example of how both 
cases could be used in the next version. Thank you.

[JMC] Additionally, a few comments:

What is iSchedule?  I see one reference to it, and it is not explained (in the 
new schedule YANG module).
[Qiufang] It's supposed to be iCalendar (Internet Calendaring and Scheduling 
Core Object Specification), as defined in RFC5545. We will fix this and add 
more clarification.
I would add values to the days of the week (perhaps use the same numeric values 
as cron).
[Qiufang] Do you mean add the "value" statement for the enumeration weekday 
type? And as cron, use 0 for Sunday, 1 for Monday...,6 for Saturday, right? 
That's a good comment. Will add in the next version.

Why is bysecond 0..60 whereas byminute 0..59.  I would think both would be 
0..59.
[Qiufang] I think that the range from 0 to 60 is because of a leap second, 
which is a one-second adjustment very occasionally applied to UTC. E.g., the 
latest leap second occurred at 07:59:59 on January 1, 2017, which is Beijing 
time: 07:59:60. RFC5545 also defines that implementations that do not support 
leap seconds SHOULD interpret the second 60 as equivalent to the second 59. 
Will explain this exception in the description.
Ultimately, this draft yearns for examples, especially of this schedule 
mechanism which, while more complete than the initial draft, brings a lot of 
complexity.
[Qiufang] Yes, the current draft defines the complete and general schedule 
mechanism which complies with recurrence rule in RFC5545, with the intention 
that this could not only be used to apply time condition based ACL policies 
here, but may also be useful for future recurrence-based schedule or event 
definition.
I agree that more examples are really needed, this is what we are working on 
now, will try to submit a new version in the near future before the submission 
window closes.


Best Regards,
Qiufang

_______________________________________________
OPSAWG mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/opsawg

Reply via email to