From vigi@icarus.UUCP Mon Nov 16 15:36:16 1992
Received: by cbmvax.cbm.commodore.com (5.57/UUCP-Project/Commodore 2/8/91)
	id AA17407; Tue, 10 Nov 92 03:55:17 EST
Received: by cbmehq.esco.so.commodore.com; Tue, 10 Nov 1992 08:49 MET
Received: by cbmita.it.so.commodore.com; Sun,  8 Nov 1992 03:01 MET
Received: by icarus.UUCP (V1.15/Amiga)
	id AA00031; Tue, 3 Nov 92 23:16:07 CET
Date: Tue, 3 Nov 92 23:16:07 CET
Message-Id: <9211032216.AA00031@icarus.UUCP>
From: vigi@icarus.UUCP (Vincenzo Gervasi)
To: cbmvax!gregm
Cc: cbmehq!cbmita!icarus!paolo
Subject: RE: Pls. Fill out this form and mail it back
Content-Type: text
Content-Length: 4983
Status: RO
X-Status: 


--0-539159445-719159714:#13860
Content-Type: TEXT/PLAIN; charset=US-ASCII



Greg Miller                         "Dad!  The heathen is getting away!"
Amiga Networking Group
gregm@cbmvax.commodore.com


--0-539159445-719159714:#13860
Content-Type: APPLICATION/octet-stream; name=asciifile
Content-ID: <Pine.2.26-experimental.9210151014.C13860@cbmvax>
Content-Description:

Request for Assignment of a SANA-II Hardware Type
Amiga Networking Group

The Amiga SANA-II Data Link standard interface utilizes specialized
identification numbers called "Hardware Types", that uniquely describe
the method of networking for individual pieces of Amiga networking hardware.



Not every piece of Amiga networking hardware needs to be given a specific
hardware type.  For instance, if company XYZ created a new ARCNET board for
the Amiga, they wouldn't need to request a new hardware type -- one already
exists for ARCNET, and all ARCNET drivers should identify themselves as such.
The same methodology applies for all Ethernet boards, all LocalTalk boards,
etc.  The drivers that do need a specific hardware type assigned are those
where the operation is entirely proprietary, and not merely a different
implementation of a standard networking protocol.  Please note that these
numbers reflect the hardware method used to communicate -- not the individual
implementation method.  As an example, a Zorro II Ethernet board,
a Zorro III Ethernet board, a PCMCIA Ethernet board, and a parallel-port
Ethernet adapter should all contain the hardware type S2WireType_Ethernet.

The following hardware types are currently assigned (from the include file "sana2.h"):

/*
** defined Hardware types
**
**  If your hardware type isn't listed below contact CATS to get
**  a new type number added for your hardware.
*/

#define S2WireType_Ethernet             1
#define S2WireType_IEEE802              6
#define S2WireType_Arcnet               7
#define S2WireType_LocalTalk            11

#define S2WireType_PPP                  253
#define S2WireType_SLIP                 254
#define S2WireType_CSLIP                255


Please fill in the following form, and return it to your Commodore
representitive.

Request for Networking Hardware Type:
						S2WireType_DyLAN

Company Name:
						ICARUS ug

Short Product Description:
	DyLAN - A low-cost, .5 Mbps network for any Amiga

If applicable, manufacturer/product numbers:

        Manufacturer Number:

        Product Number:

If the above was not applicable, describe how the product connects to the
Amiga:
	It connects through the external disk connector

Hardware Network Addresses:

        If applicable, length (in bits) of hardware network
        address:

        	We plan no hardware address...

        If applicable, is the hardware address programmable?
           (Can it, or must it be set by software?)


        If applicable, are the network addresses guaranteed to be entirely
        unique?
        	If this means "unique all around the world", no.
        	They will be unique in a single network, though.
        	(We are speaking of software-assigned numbers, aren't we ?)

Is the interface operation defined in a publicly available
standard?
		No.

If not, please include a short description of operation.  (This is not
a requirement, but is strongly advised.)

	The network hardware is based on the standard disk hardware found
	on any Amiga. It uses the disk DMA for transmission & reception, and
	several other features of the disk subsystem (i.e., syncword,
	irqs etc.) to simplify life.
	This approach has several advantages:

		- The external disk connector is the very same on every
		  model of the Amiga family;
		- By using the system-supplied hardware, we can economize
		  on the network interface hardware;
		- The throughput obtained is quite satisfactory (.5 Mbps,
		  well suited to classroom-wide networks that are our
		  primary target);

	The only disadvantage is probably that a station is (temporarly)
	disconnected from the network while accessing true drives. This is
	not a great concern in our planned environment, since we aim at a
	substantially "diskless" set of economical Amigas (e.g., 500s & 600s),
	connected to a more powerful machine (the teacher's one).

	The protocol is a custom variation on CSMA system (note: without CD !);
	we plan to have a version with two CSMA buses (allowing for two
	simultaneous transfers, up to 1 Mbps throughput) ready at some point
	in the future.

	We plan to implement broadcasts, while multicasts are a secondary
	target (most surely, they will not be present in the first release).

	The addresses (8 bits), will include a bit to indicate "escape",
	allowing both 127 machines in a single network, and a way to switch
	to a different addressing method should it be necessary. The primarly
	envisioned use of this mechanism is to indicate packets that should
	be routed through some bridge and "leave" the current network.

--0-539159445-719159714:#13860--


