Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Sun, 5 Jan 2003 05:40:05 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Sun, 5 Jan 2003 05:40:05 -0500 Received: from ns.indranet.co.nz ([210.54.239.210]:47574 "EHLO mail.acheron.indranet.co.nz") by vger.kernel.org with ESMTP id ; Sun, 5 Jan 2003 05:39:58 -0500 Date: Sun, 05 Jan 2003 23:47:58 +1300 From: Andrew McGregor To: Andre Hedrick cc: linux-kernel@vger.kernel.org Subject: Re: Gauntlet Set NOW! Message-ID: <2043540000.1041763678@localhost.localdomain> In-Reply-To: References: X-Mailer: Mulberry/3.0.0b10 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2535 Lines: 75 Oh, that's nice! Presumably you could substitute DCCP or whatever for TCP. I like it. So how about this, the result of a corridor conversation at an IETF: It is perfectly doable, using HIP and some (admittedly expensive) hardware crypto gear to run iSCSI encrypted at Gigabit Ethernet rates and faster, while being able to attach endpoints more or less at random in IP space and move them around freely while connected. Mobile hotplug IP storage :-) HIP is the Host Identity Payload, which can be seen as different things depending on which features you like. The idea starts from distinguishing the IP address, which basically represents a location in the net, from the Host Identity, which is a public key that identifies an endpoint. By some machinations, you end up being IP numbering and version agnostic, while having an extremely lightweight opportunistic key exchange protocol. There are several implementations and all the specs linked to at http://www.hip4inter.net/, not presently including my own, which is purely userspace (everything I have so far needed is provided by standard kernels, except ESP and that is now in too), BSD licensed and written in Python and which will be released soon, for some value of soon. This is a less mature protocol than iSCSI at this point, but I think there are some very interesting possibilities by combining the two. Andrew --On Saturday, January 04, 2003 21:31:39 -0800 Andre Hedrick wrote: > On Sun, 5 Jan 2003, Andrew McGregor wrote: > >> By the way, I'm principally a developer of communications standards and >> hardware, not so much software. > > I forgot to mention the template model on each side of the iSCSI protocol > state machine we have developed is agnostic? > > Initiator --- Transport --- Target --- Spindle > > TCP SCSI > Quads ATA > SCI SATA > Myrinet MD > InfiniBand LVM > TELCO USB > CARRIER 1394 > SAS > Fibre Channel > > FLOPPY, for emergencies. > > Create Your Own Create Your Own > > Yeah, I am nutter than a fruitcake, but it works! > > This is for Larry McVoy, it is the closest thing you will ever see today > which looks like a disk with an RJ-45 port. > > Cheers, > > Andre Hedrick > LAD Storage Consulting Group > > - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/