RE: [Embedlets-dev] Peer-to-peer or master/slave

[James C. <ca...@vi...>]

Hi All,

I2C of course offers peer-to-peer in with multi-master modes and
collision detection. It is the natural contender for a true peer-to-peer
protocol as it is already an industry standard.

I have no problem with Teds idea of using I2C for peer-to-peer when it
is absolutely a must. It is also a fast protocol, as fast as SPI, has a
specification for collision detection with hardware support for this.=20

I see two things emerging here

UART/SLIP for medium performance Master/Slave Client/Server polling
applications which will cover about 80% of the application space. It has
the huge advantage as Ted noticed of being able to plug into the PC
making it easy to test application etc.

I2C/SLIP for high performance peer-to-peer networking Client/Server/UDP
which allows applications to also have nodes generate asynchronous
events.

This is of course only the lowest layers of the Networking Stack so
toplevel interfaces can be na=EFve about what's going on down below.

James Caska
www.muvium.com
uVM - =91Java Bred for Embedded=92



-----Original Message-----
From: emb...@li...
[mailto:emb...@li...] On Behalf Of
Bruce Boyes
Sent: Friday, 27 June 2003 10:31 PM
To: emb...@li...
Subject: [Embedlets-dev] Peer-to-peer or master/slave


Topic tags:[ARCH][JAPL][WIRING][DOCS][MGMT][STRATEGY][NEWBIE]
_______________________________________________

At 01:01 PM 6/27/2003 -0700,=20
emb...@li... wrote:

>It is just an idea, but it will work just fine.  It will provide a=20
>network exactly like you would see with a TCP/IP switched network, in=20
>terms of topology.  Broadcast networks are fine for one or two users,=20
>but degrade quickly...  There's plenty of experience that says about=20
>20% utilization is all that you can count on before you get a dramatic=20
>degradation for each percentage above that.
>
>Clearly there is research going on that has shown some ways to make the

>degradation less dramatic (search google for 'CSMA degradation'). =20
>LonWorks has a proprietary solution as do others specific to their use.

>The easies and simplest is the round-robin, equal time, equal access=20
>mechanism that I described.  It will provide dependable, predictable=20
>data transfer from one peer to another.

...only if there is a reliable master, if I understand your idea. In a
true=20
peer-to-peer ad-hoc (I realize I added that just now) network you can't=20
count on any node being present at any time. In fact, in any realtime=20
control network I think that is a good assumption. Whatever nodes happen
to=20
be present carry on as best they can, rather than the whole network just

crashing.

So if a master is required to facilitate all communication amongst
slaves=20
there can be no true peers. Only the master and slave(s), and the
network=20
should not be called a peer network.

Bruce=20




-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
_______________________________________________
Embedlets-developer mailing list
Emb...@li...
https://lists.sourceforge.net/lists/listinfo/embedlets-developer