Re: [bmwg] Question regarding RFC 2889

[Vladimir Vassilev <[e-mail]>]

On 9/13/22 11:58, Leonard Yu wrote:
>
> Dear authors,
>
> My name is Leonard, the product manager for automated test suites at 
> Xena Networks. Our team has a debate regarding RFC 2889 Section 5.7.3, 
> and no one can persuade everybody. I think the best is to ask what you 
> think.
>
> The algorithm says that the N addresses from the Learning Port equals 
> the number of addresses supported by the DUT, aka the capacity. But 
> when the Test Port sends frames to the DUT, the DUT will also learn 
> the MAC address of the Test Port. This will make the address table 
> size N+1, since the address table is a shared resource by all the 
> ports of the DUT.
>

What about the devices that have an implementation that limits the MACs 
per port ?

 From sec. 5.7.3 Procedure:

"Using a binary search approach, the test targets the exact number of
  addresses supported per port with consistent test iterations."

/Vladimir


> We think N+1 should be the capacity (for the 3-port scenario), not N.
>
> What is your opinion regarding this? Looking forward to hearing your 
> feedback. Thanks!
>
> Best regards,
>
> Leo
>
> Leonard H. Yu
> Senior Product Manager
> Xena Networks
>
> [e-mail] <mailto:[e-mail]>
>
>
> _______________________________________________
> bmwg mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/bmwg