OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

smartgrid-interest message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [smartgrid-interest] Draft Charter - Energy Market Information Exchange

No one can really can predict how technology will evolve in the future - so I think it is good to question the assumption: will/should everything move to IP?

Last Oct I wrote a blog entry about this very topic:

The summary of that blog article was that technologies like 6LoWPAN do allow us to run IP all the way to the edge devices.  So today I do actually have the choice between IP and legacy fieldbuses. My perspective is that given the choice, to choose against IP is to bet against the Internet.

The alternative to not running IP to the edge is what has largely been done for the last twenty or so odd years.  You run one backbone network (which is definitely IP) to a gateway device.  Then you run a fieldbus network from the gateway to edge devices.  Fieldbuses include things like ZigBee, Bacnet, Lonworks, Modbus, etc - the list is endless.  I personally think there are a lot of flaws with this model:

Economies of scale: virtually everyone involved in software knows IP, often with pretty deep knowledge.  How many people on the planet understand obscure technologies like ZigBee or Modbus?  If your goal is to create open communications infrastructures, then choosing IP makes it accessible to everyone.  Choosing anything else always involves choosing a technology which only has appeal to a tiny vertical domain.

IP is Pluggable: one of the beautiful aspects of IP is that it has a clean layered model.  Once I have IP, I can run UDP or TCP.  From there I can build application protocols like HTTP.  From there I can define multiple MIME typed data formats.  This stack was designed 30 years ago and has proven to work incredibly well.  Yet every time I take a look one of these fieldbuses (almost all of them invented since TCP/IP), I continue to be amazed that they try to invent the whole stack from top to bottom in a proprietary, closed fashion.

IP is Scalable: IP runs the Internet - there is no network which has proven itself to be more scalable.  It was designed from the ground up to be *the* inter-networking protocol, and does this remarkably well.  Fieldbuses on the other hand are designed explicitly to be islands of networking.  They all end up implementing proprietary addressing, routing, and data formats.

Gateways vs Routers: when you run IP from end-to-end on your networks, then the nodes in the middles are only routers.  Gateways on the other hand have to translate between data formats and often data models.  Translating between data models always results in impedance mismatch and information degradation.  Edge devices can't be globally addressed directly, and you are at the mercy of the gateway for communicating data, provisioning, and network management.  This is always a sub par solution compared to communicating directly with the edge device.

I've spent my career working on this very problem: how to map fieldbuses to Internet technologies.  And I don't think the architecture of yesterday works well.  I think IP to the edge is an opportunity to architect our systems to be more open, reliable, and flexible.  That is why I am such a huge believer in 6LoWPAN - and not just for wireless, but its techniques work equally well over 485 and other fieldbus types of media.


On Thu, Apr 2, 2009 at 1:40 AM, Michel Kohanim <michel@universal-devices.com> wrote:

Hi Brian, I had to think about this for a day! Perhaps my statements are going to cause a heated debate, but, here I go:


I really do not see why everything has to communicate IP. What’s wrong with having devices communicate in their own native languages and over their desired (most optimal) media? What do we gain by having all devices communicate IP if – and as you (Brian) suggested – we do not first come up with the abstract model? We have a hammer?

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]