search

dhcwg / yang

A workplace for IETF drafts related to DHC WG and YANG
1 stars 1 forks source link

hackathon101: vendor-info element is mandatory. it shouldn't be #22

Open tomaszmrugalski opened 6 years ago

tomaszmrugalski commented 6 years ago

This is the first of several issues and improvement suggestions that are a result of the IETF Hackathon 101 in London. The goal of our team was to extract the YANG model from the latest -06 version, load it into sysrepo, then apply configuration and then retrieve it from a Kea-netconf daemon, a new daemon that was started during the hackathon. This issue will be updated with a link to the hackathon summary.

The vendor-info element in server-config/server-attributes in ietf-dhcpv6-server model is mandatory. It shouldn't. All modern servers allow to configure special options for certain vendors, it is never mandatory to configure any of them. Also, most reasonable servers are able to configure options for multiple vendors.

zihaohe123 commented 6 years ago

Got your point, thanks. BTW, since 'vendor-info' is non-mandatory, it makes wonder whether 'DUID' should be mandatory. According to 3315bis, 'DUID' is more like something that is generated, rather than configured (of course it could be configured).

bevolz commented 6 years ago

It depends where.

For recording actual leases assigned by a server based on received client request, yes. (And for client model when describing client.) But in probably other places, no. For example, a client reservation should not require duid as other things can be used to identify a client (such as a port on a switch & the switch) — but once this turns into a lease, duid (client-id) for actual client must be recorded (it could change though if different client plugged into that port). This is needed for Reconfigure & and reporting. At least, that’s the model that would be required for the server I work on.

On Mar 26, 2018, at 7:23 AM, Zihao He notifications@github.com<mailto:notifications@github.com> wrote:

Got your point, thanks. BTW, since 'vendor-info' is non-mandatory, it makes wonder whether 'DUID' should be mandatory. According to 3315bis, 'DUID' is more like something that is generated, rather than configured (of course it could be configured).

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/dhcwg/yang/issues/22#issuecomment-376133033, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ABm3YQBw607ZHKHGy_Or0cqbdhDbreBLks5tiM-1gaJpZM4SwMqx.