From owner-freebsd-net Wed Nov 15 8:34:47 2000 Delivered-To: freebsd-net@freebsd.org Received: from spider.pilosoft.com (p55-222.acedsl.com [160.79.55.222]) by hub.freebsd.org (Postfix) with ESMTP id 7841737B4C5 for ; Wed, 15 Nov 2000 08:34:43 -0800 (PST) Received: from localhost (alexmail@localhost) by spider.pilosoft.com (8.9.3/8.9.3) with ESMTP id LAA05586; Wed, 15 Nov 2000 11:35:32 -0500 (EST) Date: Wed, 15 Nov 2000 11:35:32 -0500 (EST) From: Alex Pilosov To: Julian Elischer Cc: "Gary T. Corcoran" , freebsd-net@FreeBSD.ORG Subject: Re: netgraph/atm In-Reply-To: <3A126A2D.7F43B120@elischer.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Wed, 15 Nov 2000, Julian Elischer wrote: > My question is whether there are any boards that can give us the > physical interface to a DSL line so that we can run this ATM code over > a DSL line? It's one thing having the software but the bits need to > get out of the system somehow. I'm looking at the ISP side of this. To ISP, DSL service is always delivered as normal ATM circuit. For customer's side, its also delivered as low-speed atm connection. There is already a driver for Efficient Networks hardware "lanai" by someone in .ee, which the author plans to netgraph-enable too. > One thing I just thought about, is that if we are simulating an > ethernet interface and bridging the packets using mac/llc > then wouldn;t we require that we have some unique MAC address > for that virtual interface so that things don't break? > (can we just always use broadcast?) are the packets multiplexed > using destination MAC address? True. Cisco in this case generates random number as a mac address. For us, we must add support for SIOCGIFADDR to 'virtual ethernet' device (if its not already there). In case of broadcast, it should be rebroadcast to all VCs in the bridge group, if that's what you are asking. > Also which of the two ATM stacks in FreeBSD is the most suitable? > I'm guessing HARP, but NATM, because it is such a very basic > framework, might be suitable for netgraphification.. I don't think it would be very hard to netgraph-enable HARP. But we'll see, once I start writing code ;) > (I have no idea if I'm talking rubish but > a simple netgraph node that does aal1/aal5 > would fit into the netgraph world very well.) There really isn't a need to do AAL1/5, since its always handled by the ATM card in hardware (you just have to tell it to do it). The cool part is extraction of specific VC traffic as netgraph streams. The very-very ambitious stuff is to have SCSP node (SVC establishment protocol) as a node in netgraph, and reorganize all the HARP code to be nodes and edges, handling routed packets via netgraph as well... But this will require a lot of rethinking, and semantics for routed interface is unclear...At any case, this is stuff for later ;) -alex To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message