lwip-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[lwip-users] RE: [lwip] Deploying Ethernet MAC addresses


From: Jason Morgan
Subject: [lwip-users] RE: [lwip] Deploying Ethernet MAC addresses
Date: Wed, 08 Jan 2003 22:53:10 -0000

Hi,

OUI's are far from free, I think last time we bought one it was about $1500
dollars per 16m addressed.

They do have a 'low cost' service that allocates 65,000 instead of
16,000,000 addresses.

I believe that there are also ranges set aside for special use, VLANS and
PPP, I suppose you could
use one of these for testing.

Personally, for software testing I have some old cards, the probability of
any of their allocated range
appearing on our LAN is so small (=zero) that we can destroy one of these
cards and use its MAC or OUI
safely, obviously not for production though. The company that made these
cards ceased to exist some time ago.

Less safely, just guess a number, or base one on the above 'stolen' OUI.

We also have a range of MAC addresses available from our own OUI that are
reserved for engineering use only, trouble is, that the probability of two
engineers using the same one is actually higher, given that this range is
only 255 numbers.

At one time Dallas had a uniqueware device that allocated you a number on a
per-chip bases, but I think they fell out with the IEEE over this and
obsoleted it, now you have to own the OUI.

The OUI rules (available on the IEEE website) state that you must record the
use of each MAC and that you must have records of using 90% of them before
you can ask for more, failing to do this can loose your OUI.

Oh, you don't need to be a corporate member of the IEEE to allocate then,
they are just the controlling body.


Regards,

Jason.

P.S.  Disclaimer: I or my employer do not condone using other companies or
made up OUIs or MAC address for testing or any other purpose, if you mess up
its your fault!!

-----Original Message-----
From: address@hidden [mailto:address@hidden
Sent: 17 December 2001 14:53
To: address@hidden
Subject: [lwip] Deploying Ethernet MAC addresses



Hello Adam and others,

if you plan on deploying Ethernet interfaces in the real
world (i.e. outside your test environment), using a
unique MAC address is a must. See
     http://standards.ieee.org/regauth/oui/tutorials/
and for our (Ethernet) purpose, specifically:
     http://standards.ieee.org/regauth/oui/tutorials/lanman.html

This is where the list of OUI (organizational unique id) are:
     http://standards.ieee.org/regauth/oui/index.shtml

I'm not aware of a public OUI or MAC address is available for
testing. Even 00-00-00 is already given away (to Xerox).

I believe requesting a OUI range is free (or maybe needs
IEEE membership?), but it requires you to administer
every given MAC you put in the real world, AFAIK.

The links above should give all the details. Regards,

Leon.



 

                    Adam Dunkels

                    <address@hidden>       To:     address@hidden

                    Sent by:             cc:

                    address@hidden       Subject:     Re: [lwip] Ethernet
MAC address now 3 x u16_t?                
                    s.se

 

 

                    17-12-01 15:00

                    Please respond

                    to lwip

 

 





Hi Leon!

On Monday 17 December 2001 11.55, you wrote:
> what's the rational for changing the MAC address struct to
> u16_t?

The only reason was that I was testing out the 16-bit char stuff that was
discussed here a while ago. Also, 16-bit accesses are a bit faster than
8-bit
on most CPUs (or am I wrong here?).

> I think six "u18_t" are the obvious choice, given that:
> - a MAC address is defined to be 6 octects
> - is split in two halves of six u16_t
>   (of which three organization unique octets)

You definately have a point here. So you recommend going over to u8_t
again?

By the way, aren't there some kind of private MAC-addresses? I guess such
an
address should be the "default" address in the lwIP distribution files.

> BTW, the cs8900a driver is ready to be released RSN,
> lwIP 0.5 compatible and tested (somewhat).

Great!

/adam
--
Adam Dunkels <address@hidden>
http://www.sics.se/~adam
[This message was sent through the lwip discussion list.]




[This message was sent through the lwip discussion list.]
[This message was sent through the lwip discussion list.]




reply via email to

[Prev in Thread] Current Thread [Next in Thread]