Re: proposed example text and question on draft-ietf-rtgwg-policy-model

[Chris Bowers <[e-mail]>]

Acee,

The existing text provides a clear description of import-policy and
export-policy.  It uses the terms import and export from the point of view
of the RIB.  I think it is important for the terminology to remain
consistent.

Chris
======

       leaf-list import-policy {
         type leafref {
           path "/rt-pol:routing-policy/rt-pol:policy-definitions/" +
             "rt-pol:policy-definition/rt-pol:name";
           require-instance true;
         }
         ordered-by user;
         description
           "List of policy names in sequence to be applied on
            receiving a routing update in the current context, e.g.,
            for the current peer group, neighbor, address family,
            etc.";
       }

=============

    leaf-list export-policy {
         type leafref {
           path "/rt-pol:routing-policy/rt-pol:policy-definitions/" +
             "rt-pol:policy-definition/rt-pol:name";
           require-instance true;
         }
         ordered-by user;
         description
           "List of policy names in sequence to be applied on
            sending a routing update in the current context, e.g.,
            for the current peer group, neighbor, address family,
            etc.";
       }

==========


On Wed, Jun 24, 2020 at 2:24 PM Acee Lindem (acee) <[e-mail]> wrote:

> Hi Chris,
>
>
>
> *From: *Chris Bowers <[e-mail]>
> *Date: *Wednesday, June 24, 2020 at 2:05 PM
> *To: *Yingzhen Qu <[e-mail]>
> *Cc: *Acee Lindem <[e-mail]>, Chris Bowers <[e-mail]>, "
> [e-mail]" <
> [e-mail]>, Routing WG <[e-mail]>
> *Subject: *Re: proposed example text and question on
> draft-ietf-rtgwg-policy-model
>
>
>
> Yingzhen,
>
>
>
> I think the following changes in naming are clearer:
>
>
>
> set-import-level -> set-export-level
>
> import-level -> export-level
>
>
>
> IS-IS is the protocol where it ultimately is advertised so it is *imported
> *here and NOT from the protocol from which is is *exported.* Between
> import and export, the former is clearly cleaner.
>
>
>
> I understand that the model supports both import and export policies.
> However, as far as I can tell, 'isis-level-2' should never be used in an
> import policy, only an export policy.   Instead, 'isis-level-2-type' would
> be used in an import policy.  The name change that I propose above makes
> this clear.
>
>
>
> Acee's proposal to use 'set-level' for isis-level-2 leaves this unclear.
>
>
>
> See above. *set-isis-level* is probably better than * set-level. *In any
> case, we aren’t going to change as you suggest.
>
>
>
> Thanks,
> Acee
>
>
>
>
>
> Thanks,
>
> Chris
>
>
>
> On Fri, Jun 19, 2020 at 12:08 AM Yingzhen Qu <[e-mail]>
> wrote:
>
> Hi Acee and Chris,
>
>
>
> I will change the name in next revision with other comments.
>
>
>
> Thanks,
>
> Yingzhen
>
>
>
> *From: *"Acee Lindem (acee)" <[e-mail]>
> *Date: *Thursday, June 18, 2020 at 4:21 PM
> *To: *Yingzhen Qu <[e-mail]>, Chris Bowers <
> [e-mail]>, "[e-mail]" <
> [e-mail]>, "[e-mail]" <[e-mail]>
> *Subject: *Re: proposed example text and question on
> draft-ietf-rtgwg-policy-model
>
>
>
> Hi Yingzhen,
>
>
>
> Meant to reply earlier. Thanks for responding.
>
>
>
> *From: *Yingzhen Qu <[e-mail]>
> *Date: *Thursday, June 18, 2020 at 6:59 PM
> *To: *Chris Bowers <[e-mail]>, "
> [e-mail]" <
> [e-mail]>, Routing WG <[e-mail]>
> *Subject: *Re: proposed example text and question on
> draft-ietf-rtgwg-policy-model
> *Resent-From: *<[e-mail]>
> *Resent-To: *<[e-mail]>, Jeff Tantsura <
> [e-mail]>, Acee Lindem <[e-mail]>, Xufeng Liu <
> [e-mail]>
> *Resent-Date: *Thursday, June 18, 2020 at 6:59 PM
>
>
>
> Hi Chris,
>
>
>
> Thanks for the review and proposed examples, really appreciate.
>
>
>
> I’ve uploaded a new version of the draft and included the example to
> demonstrate route redistribution between ospf and isis. I didn’t include
> the one to install ospf routes to RIB considering this is default behavior
> unless you specify a policy to limit the ospf routes installation.
>
>
>
> Regarding the name, the model supports both import and export modes, so I
> didn’t want to simply change the name to “set-export-level”, but open to
> suggestions. The model also provides a grouping “apply-policy-group” that
> can be used by routing protocols for route redistributions, and there are
> descriptions about it in Section 6.
>
>
>
> I Think we should change it to set-isis-level or simply set-level.
>
>
>
> Thanks,
> Acee
>
>
>
>
>
> Thanks,
>
> Yingzhen
>
>
>
> *From: *Chris Bowers <[e-mail]>
> *Date: *Wednesday, June 10, 2020 at 2:23 PM
> *To: *"[e-mail]" <
> [e-mail]>, "[e-mail]" <[e-mail]>
> *Subject: *proposed example text and question on
> draft-ietf-rtgwg-policy-model
> *Resent-From: *<[e-mail]>
> *Resent-To: *<[e-mail]>, <[e-mail]>, <
> [e-mail]>, <[e-mail]>
> *Resent-Date: *Wednesday, June 10, 2020 at 2:23 PM
>
>
>
> I would like to propose adding the following example to the text of
> draft-ietf-rtgwg-policy-model
> to better illustrate how the model will work in practice with routing
> policies involving IGPs.
> The proposed text is shown below.
>
>
> I think that the example below also illustrates a problem with the naming
> of what is currently called "import-level" and "set-import-level".  In the
> example, the export policy called
> "export-all-OSPF-prefixes-into-ISIS-level-2" uses the "set-import-level"
>  action.  As far as I can tell, it only makes sense to use
> "set-import-level" in an export policy, and not in an import policy.  If
> this is the case, wouldn't it make more sense to call it "set-export-level"?
>
>
> ===========
>
> Proposed text for new IGP routing policy example:
>
>
>
> This example illustrates the import and export policies corresponding to
> the following scenario.
>
> All routes that are learned via OSPF advertisements should get installed
> in the RIB.
>
> All routes in the RIB that have been learned from OSPF advertisements
> corresponding to
>
> OSPF intra-area and inter-area route types should get advertised into ISIS
> level 2 advertisements.
>
>
>
>           <policy-definitions>
>
>            <policy-definition>
>
>              <name>import-all-OSPF</name>
>
>              <statements>
>
>                <statement>
>
>                  <name>term-0</name>
>
>                  <conditions>
>
>                    <match-prefix-set>
>
>                      <prefix-set>all-prefixes</prefix-set>
>
>                    </match-prefix-set>
>
>                  </conditions>
>
>                  <actions>
>
>                    <policy-result>accept-route</policy-result>
>
>                  </actions>
>
>                </statement>
>
>              </statements>
>
>            </policy-definition>
>
>               <policy-definition>
>
>              <name>export-all-OSPF-prefixes-into-ISIS-level-2</name>
>
>              <statements>
>
>                <statement>
>
>                  <name>term-0</name>
>
>                  <conditions>
>
>                    <match-prefix-set>
>
>                      <prefix-set>all-prefixes</prefix-set>
>
>                    </match-prefix-set>
>
>                    <match-route-type>
>
>
> <proto-route-type>ospf-internal-type</proto-route-type>
>
>                    </match-route-type>
>
>                  </conditions>
>
>                  <actions>
>
>                    <set-import-level>
>
>                      <import-level>isis-level-2</import-level>
>
>                    </set-import-level>
>
>                    <policy-result>accept-route</policy-result>
>
>                  </actions>
>
>                </statement>
>
>              </statements>
>
>            </policy-definition>
>
>          </policy-definitions>
>
>
>
> ==========
>
>
>
> Thanks,
>
> Chris
>
> _______________________________________________
> rtgwg mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/rtgwg
>
>