EtherNet/IP network addressing

What is the understood standard for Class network ID used for Industrial EtherNet/IP addressing for PLCs and HMI networks? A B or C?
Is there a standard range used in that Class for EtherNet/IP industrial networks? Any info on ranges would be great I have 15 nodes on one segment linked by 7 switches. Any address Info or help would be great. I have read and understand octets and differences of the classes but I can't find a standard for deriving what class and ranges explicitly used for industrial networks

This Site has all the info that I know of.

http://www.controlnet.org/02a_abenip-cn/index.htm

This gives the standard for implementing features of EtherNet/IP hardware but I have found nothing regarding standards of actual IP class specifications to a control network or range of IP static addresses.

I am not sure if I can refer to any standards but all the manufacturers of the industrial automation usually recommend using 192.168.x.x pattern with 255.255.0.0 subnet mask.

This, of course, is good for a separate network, not connected to anything else, like a corporate net. In that case, since most of these devices require a fixed IP address, you would have to use whatever the network administrator will let you to.

However, it is a much better idea to have all your stuff on the separate net and, if a connection to the outside world is required, to use a router. In this case, your automation net will not be affected by the corporate traffic...

That addressing makes this a Class B network scheme with only 2 octets set and mask at 255.255.0.0 but the problem that I see with that is it falls into the addressing of a class "C" 192-223 as min max octet addressing? I don't get how they can recomend this when it falls out of conventional archetecture for Ethernet standards. This isn't making a whole lotta sense to me. Even the Allen bradley manual confuses me with there examples from "SLC 500 Instruction Set" manual you see example of addressing page 13:31, Table 13:10. It shows address' of 12.34.56.1 which is a class "A" network address "A" 1-127. Then further down they show another example on page 13:35 utilizing a 130.151.194.19 IP address which is a class "B" 128-191. Somebody help!!! I want to do this in an orthodox professional manner that coresponds to a standard!!

I don't have that Allen-Bradley manual in front of me but, as far as I remember, they simply explain the concept of sub-netting to those people who have never heard of it. Therefore, all the IP addresses in those examples are pure fiction, for illustrational purposes only.

GE Fanuc illustrates its manuals with addresses like 3.0.0.1 This is a Class A net and (if I am not mistaken) anything that begins with 3. has been reserved for General Electric Corporation. Can you use these without their permission?

I think the correct and professional classification of your network has any meaning only if you are talking about the Big Internet. This is exactly what you don't want to do with your industrial devices: have them exposed directly to the Net.

As I have mentioned above, your system might be connected to the corporate or plant network - but not to the Internet. Therefore, you will either have to comply with the plant rules - or to use whatever you feel like using, regardless of the class.

Once again, the most common addresses I have seen are 192.168.1.1 and up.

Good Info

Read This

We use, and I recommend, 192.168.0.0 for local networks. 192.168.0.0 is reserved for local networks. This range of IP addresses should never be on the internet.

If you are just going to network your PLC's to themselves and a few peer to peer PC's that are not on the corporate network use the 192.168.0.X with a subnet of 255.255.255.0. This gives you 254 addresses to play with and it meets the convention that everybody uses.
When I setup my PLC ethernet network I read the AB CLX manual and they used 130.130.130.x for IP addresses. Thinking "Well this is what you should use", I setup all my PLC's in that range. THEN I read alot of what they other guys are telling you. Now I need to go back and change them to the 192.168.0.x but that is going to be a pain in the neck.

OK so if there is no TRUE standard how about I make the Value for the 3rd and the 4th octet represent equipment ID. Say a piece of equipment has a PLC and HMI on Ethernet. I creat my IP address for that piece of equipment as 192.168.1.0 for the PLC and 192.168.1.1 for the HMI. The next machine would be 192.168.2.0 for the PLC and 192.168.2.1 for the HMI. The network ID would then be 192.168.X.X with asubnet mask of 255.255.0.0.
Would that be an acceptable scheme.
Any other ideas along those lines would appreciated.

Plan ahead for link to corporate network

In our corporation, the engineering and IT departments agreed that the fixed IP addresses in PLC's and OIT's would be set to values that MUST be outside any other addresses that are now or ever will be used by the enterprise network.

Two reasons for this: 1) We are sure that EVENTUALLY most of the OIT's will be on the plant network (through switches, of course) to allow data collection, component verification, etc. 2) If we had an OIT with a "safe" address and a PLC with a "dangerous" address (ie, same as a desktop or server somewhere in enterprise) someday, someone will get the Ethernet connectors confused and accidently put the PLC on the enterprise network. If its address is identical to some server or desktop somewhere... big trouble!! Our greatest concern is not that we take down the desktop, but that somehow this cause unintentional changes in PLC memory with the risk of affecting equipment operation.

Here is another site that may have some information to help you out. I think I got this one from Peter at some time or another.

Click here

Hi rkduet

Beyond dynamic/static IP addressing, a key point you should read about remains to be the difference between routable and non-routable IP addresses.

Basically routable addresses can be used to route on the Internet and non-Routable addresses cannot be used to route on the Internet. That's why non-routable IP addresses are usually used behind firewalls.

That said you're free to assign as desired your IP addresses on a private network (even by using routable IP ranges as long as they won't conflict with dynamic addresses on a public network).

Three ranges of non-routable ranges exist :
10.x.x.x
192.168.x.x
172.16.x.x
You see why many manufacturers often recommend 192.168.bla bla.
As for me I use 10.x.x.x behind routers since it lets you more usable ranges due to subnet masks.

Be sure you perfectly understand how your subnet mask operates, and allow you to build several independent parts in you network with some stations assigned as gateways i.e. Usually low numbers among available ranges are reserved for gateways/routers. Moreover, using addresses terminated by 0 is not really common.

Router 10.0.0.1 - mask 255.255.255.224

Network part 1
Station 1 - 10.0.0.2 - mask 255.255.255.252
Station 2 - 10.0.0.3 - mask 255.255.255.248

Network part 2
Station 3 - 10.0.0.5 - mask 255.255.255.224
Station 4 - 10.0.0.24 - mask 255.255.255.224

All stations can access the router
St1 can access : router + St2
St2 can access : router + St1 + St3 (and not St4)... if I'm not mistaken (what happens sometimes with such a mess and since I'm programming a PLC in the meantime )

"Clear informations" on those subjects are usually provided with your favorite OS package and mostly with MS Win xx Server or Win 2K packages. See in your company if such a platform is running... and approachable.

Hope this helps

Laurent

Thanks glaverty, That answers a lot, for I have not seen the private address ethernet standard broken down just the public. Now I see where the 192.168.0.0 is comming from. This is starting to make more since. Knowledge is power. Thanks guys, for sharing yours.