search

tsaad-dev / te

IETF TE Tunnels YANG models
16 stars 19 forks source link

Bandwidth in uint64 or bandwidth-scientific-notation #221

Closed italobusi closed 1 year ago

italobusi commented 1 year ago

The bandwidth profile parameters (cir, eir, cbs, ebs) are defined as uint64

Should they be re-defined using the bandwidth-scientific-notation?

The units for cbs, ebs are Kbytes: should we generalize the bandwidth-scientific-notation with number-scientific-notation?

tsaad-dev commented 1 year ago

Which YANG models are using scientific notation? Why do we need it?

italobusi commented 1 year ago

The bandwidth profile parameters (cir, eir, cbs, ebs) are defined as uint64

Should they be re-defined using the bandwidth-scientific-notation?

The units for cbs, ebs are Kbytes: should we generalize the bandwidth-scientific-notation with number-scientific-notation?

See past discussion in #116 and https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-client-signal-yang/issues/7

The solution (bandwidth-scientific-notation as defined in RFC8776-bis I-D) solves the problem with CIR, EIR and PIR but not for CBS, EBS and PBS attributes ...

italobusi commented 1 year ago

2023-06-23

Agreements:

  1. to have a consistent representation of packet bandwidth:

               type uint64;
           units "bits/second";

  2. remove bandwidth-scientific-notation data type

  3. to have a consistent representation of burst size

               type uint64;
           units "bytes";