From: Geoff_Thompson@BayNetworks.COM (Geoff Thompson)
To: John_Hart@3mail.3Com.COM
Cc: 100271.522@compuserve.com, P8021@nic.hep.net, langille@nexen.com,
        Paul_Frantz@BayNetworks.COM, alonge_ken@po.gis.prc.com
Subject: Re: Re: Re: REVISION TO SDE
Date: 27 Nov 1995 02:02:59 GMT

John-

I don't think I agree with you on this one.  I believe that the basis for the
original proposal to utilize 802.10 style tagging for VLAN was that 802.10
had already solved the problem in the encapsulation domain.  The reason that
they had solved it was that security customers have long required VLANs, long
before they became a hot buzz word in the bridging community.  Every LAN
security/encryption situation that I have ever run into required that you be
able to separate users into a number of separate "need-to-know" communities.

Given that the capability is already there in an 802 standard it seems silly
to me to reinvent it.  This does not mean that I am against other types of
VLAN assignment or the work for a standard to be able to traverse VLAN types,
just that we should utilize the work that has already been done.

Further I believe that, as the market matures, there is a strong liklihood
that there will be a desire to encrypt VLANs with a different key for each
VLAN.  Using 802.10 would position us for that ideally.  I don't think we can
put it in at the beginning because the market doesn't seem ready for it yet.

Geoff Thompson (My opinions, perhaps, but not necessarily those of my
employer or my Working Group)