Re: [Ntp] Antw: Re: Antw: Re: Calls for Adoption -- NTP Extension Field drafts -- Four separate drafts

[Steven Sommars <[e-mail]>]

Ethernet MAC addresses may be globally or locally unique; see "Ethernet LG
bit".
The same locally assigned Ethernet address may be used in hosts on
different L2 networks.

Each equipment cabinet in a project I was involved with used the same set
of local(internal to frame) MAC addresses.
Globally unique MAC addresses were unavailable.  (NTP was used.)
This worked fine, since the Ethernet layer 2 was confined within the
cabinet.




On Mon, Sep 16, 2019 at 5:35 PM Mark Atwood <[e-mail]> wrote:

> On Sun, Sep 15, 2019, at 23:36, Hal Murray wrote:
> >
> > [e-mail] said:
> > >> Can we assume that every server will have an Ethernet host address?
> > > The vast majority will, but I'm not sure we can rely on them being
> random.
> >
> > They are definitely not random.  They are unique.  (unless somebody
> screws up)
> >
> > Within a batch of Ethernet cards, they will probably be sequential.
> (I'll say
> > more if anybody wants.)
>
> In a past life, I've written the scripts that generate the mapping of
> manufacturing serial number to ethernet MAC addresses, and then generated
> the text files in the format requested by various manufacturing vendors so
> that they can fuse the MAC address into the device at manufacturing time,
> and so that the manufacturer and packaging vendors can print the necessary
> pcb labels and external case and box labels and barcodes.    And I've seen
> a lot of such other files that are sent to similar manufacturers.
>
> I've never seen one that wasn't sequential.
>
> The sequential relationship is so strong that it's used to identify supply
> chain and logistics glitches.
>
>
> On the other other hand, you must not bet your life or your security on
> them being truly globally unique.   There are lots of chips in the supply
> chains and in warehouses that were manufactured off-label, grey market,
> unlicensed, and off-shift.   And anyone with kernel level or UEFI write
> level access to a box can persuade a chip or it's driver to lie about it's
> MAC.
>
> ...m
>
> _______________________________________________
> ntp mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/ntp
>