Re: [Lsvr] Need clarification on IGP-Metric TLV for LS Link Attributes in BGP-SPF deployments

[Pushpasis Sarkar <[e-mail]>]

Hi Acee,

Thanks a lot for taking care of this.

Best regards
-Pushpasis


On Tue, Mar 24, 2020 at 8:18 PM Acee Lindem (acee) <[e-mail]> wrote:

> Hi Pushpasis, et al,
>
> This is clarified in the -08 version of the draft.
>
> Thanks,
>
> Acee
>
>
>
> *From: *Pushpasis Sarkar <[e-mail]>
> *Date: *Sunday, March 22, 2020 at 11:43 AM
> *To: *Acee Lindem <[e-mail]>
> *Cc: *"[e-mail]" <[e-mail]>,
> "[e-mail]" <[e-mail]>
> *Subject: *Re: Need clarification on IGP-Metric TLV for LS Link
> Attributes in BGP-SPF deployments
>
>
>
> Hi Acee,
>
>
>
> Sure. If anyone doesn't prefer any other way we can proceed with this.
>
>
>
> Thanks
>
> -Pushpasis
>
>
>
>
>
> On Sat, Mar 21, 2020 at 7:49 PM Acee Lindem (acee) <[e-mail]> wrote:
>
> Hi Pushpais,
>
> I think we can get away with this w/o modifying RFC 7752 since, as you
> noted, it just says it is variable.
>
> Thanks,
>
> Acee
>
>
>
> *From: *Pushpasis Sarkar <[e-mail]>
> *Date: *Saturday, March 21, 2020 at 10:15 AM
> *To: *Acee Lindem <[e-mail]>
> *Cc: *"[e-mail]" <[e-mail]>,
> "[e-mail]" <[e-mail]>
> *Subject: *Re: Need clarification on IGP-Metric TLV for LS Link
> Attributes in BGP-SPF deployments
>
>
>
> Hi Acee,
>
>
>
> I thought RFC7752 says it is a variable length TLV. 1-byte, 2-byte, and
> 3-byte metrics are possible lengths. Can't we add a 4th possible length via
> this draft? If not then a new TLV may be a better option.
>
>
>
> However if I remember correctly, there is a RFC7752 bis being worked upon.
> Is there an option we can add it in that document? Not sure if it will be a
> proper way or not. Just thinking out loud :)
>
>
>
> Thanks
>
> -Pushpasis
>
>
>
>
>
> On Sat, Mar 21, 2020 at 7:17 PM Acee Lindem (acee) <[e-mail]> wrote:
>
> Hi Pushpasis,
>
> If we use a length of 4, we’d need a new TLV (or modify RFC 7752). I think
> the former would be a better option.
>
> Thanks,
>
> Acee
>
>
>
> *From: *Pushpasis Sarkar <[e-mail]>
> *Date: *Saturday, March 21, 2020 at 9:00 AM
> *To: *Acee Lindem <[e-mail]>
> *Cc: *"[e-mail]" <[e-mail]>,
> "[e-mail]" <[e-mail]>
> *Subject: *Re: Need clarification on IGP-Metric TLV for LS Link
> Attributes in BGP-SPF deployments
> *Resent-From: *<[e-mail]>
> *Resent-To: *Keyur Patel <[e-mail]>, Acee Lindem <[e-mail]>,
> Shawn Zandi <[e-mail]>, Wim Henderickx <
> [e-mail]>
> *Resent-Date: *Saturday, March 21, 2020 at 9:00 AM
>
>
>
> Hi Acee,
>
>
>
> My personal preference will be having it as a 4-byte metric due to ease of
> implementation as well as encoding/decoding efficiency due word-size
> alignment. It may not be a big thing but encoding/decoding a 3-byte value
> to/from byte stream is few more CPU instructions than just a 4-byte read
> from a memory address.
>
>
>
> Thanks
>
> -Pushpasis
>
>
>
> On Sat, Mar 21, 2020 at 12:10 AM Acee Lindem (acee) <[e-mail]>
> wrote:
>
> Hi Pushpasis,
>
> I think for BGP-LS SPF we should always use 3 octet metrics. This will
> offer the most flexibility w/o redefining the TLV. If you agree, I will
> update the SPF draft to state this.
>
> Thanks,
>
> Acee
>
>
>
> *From: *Pushpasis Sarkar <[e-mail]>
> *Date: *Monday, March 9, 2020 at 11:31 AM
> *To: *"[e-mail]" <[e-mail]
> >
> *Cc: *"[e-mail]" <[e-mail]>
> *Subject: *Need clarification on IGP-Metric TLV for LS Link Attributes in
> BGP-SPF deployments
> *Resent-From: *<[e-mail]>
> *Resent-To: *Keyur Patel <[e-mail]>, Acee Lindem <[e-mail]>,
> Shawn Zandi <[e-mail]>, Wim Henderickx <
> [e-mail]>
> *Resent-Date: *Monday, March 9, 2020 at 11:31 AM
>
>
>
> Hi Authors,
>
>
>
> I need a small clarification on how the Link IGP-Metric TLV (type 1095)
> for the links originated by an BGP-SPF speaker look like. My doubt is
> specifically on what would be the length of the metric value. For example,
> following is the excerpt from RFC7752 section 3.3.2.4 which specifies the
> length to be 1, 2 or 3 bytes for ISIS narrow-metrics, OSPF and ISIS
> wide-metrics.
>
>
>
> 3.3.2.4.  IGP Metric TLV
>
>
>
>    The IGP Metric TLV carries the metric for this link.  The length of
>
>    this TLV is variable, depending on the metric width of the underlying
>
>    protocol.  IS-IS small metrics have a length of 1 octet (the two most
>
>    significant bits are ignored).  OSPF link metrics have a length of 2
>
>    octets.  IS-IS wide metrics have a length of 3 octets.
>
>
>
>       0                   1                   2                   3
>
>       0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
>
>      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>
>      |              Type             |             Length            |
>
>      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>
>      //      IGP Link Metric (variable length)      //
>
>      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>
>
>
>                      Figure 21: IGP Metric TLV Format
>
>
>
> What should be the length of the metric field when the origin is a BGP-SPF
> speaker?
>
>
>
> Looking forward to your clarification on this. Also it will be appreciated
> a lot if a sentence or two can be added to the draft clarifying the above
> in the next version.
>
>
>
> Thanks and regards,
>
> -Pushpasis
>
>
>
>