Using dynamic IP address with RSView, CLX and SLC505...

[TConnolly]

Does anyone have any experience with using dynamic IP addresses (bootP) for 1756-ENBT modules, SLC5/05s, and RSView?

If the dynamic IP address is changed how does RSView find the right node again?

Maybe this is simple and I'm overlooking something.

Right now I use static IP addresses for our PLCs, but we are running out of allocated static IP addresses. I have just installed a new ControlLogix gateway to bridge to PLC5s and SLC5/04s. If I change to using dynamic IP addresses, what kind of problems can I expect?

I already know what kind of problems to expect if I go to theIT trolls for more static addresses.

 

[TimWilborne]

If you look through the documentation of most of the ethernet devices, it says if you use BOOTP or DHCP it should assign the same address everytime. I think you will run into more headaches doing this than you will going to the IT trolls.

 

[plchacker]

If you use BootTP, you have to set up a BootTP server. The server looks for the MAC address (in hex, on the device, it can not be changed) of each device, and then asigns a IP address acording to the specs in the server. From what I can gather, BootTP is AB's version of DHCP. The difference is that you have control of it, not the other way around.


With SLC505, and PLC5XXE, it is fairly simple to set static IP addresses. You get online with a DH+ or RS232, and set the IP address. This changes with the CLX Ethernet Module, 1756 ENBT. Its been over a year, and I'm a bit foggy, but if I remember right, you must first contact the IP module through Ethernet, in order to set it up. BootTP makes this process much easier. Like I said, its been over a year. If you follow the instructions that comes with the module, you can step right through the setup.

For RSview, you use the assigned IP address (through RSLinx). It should always be the same once it is set up.

In general, I would avoid the IT people when it comes to your control network. If that is not possible, be polite, but be stern about the needs of the control network. Ask for a range of IP addresses, and promise to keep them informed of the use of these addresses. Then you can pick and choose within that range. I organized IP address according to process location in the mill. Just a thought.

 

[Gerry]

BOOTP and DHCP are different (I'm sure Alaric knows this).

BOOTP assigns a fixed IP address to a device based on the device's MAC address. The correlation is set up in a table in the BOOTP server. The device will always be assigned the same IP address until someone changes the server's tables.

DHCP assigns a dynamic IP address from a pool of available addresses maintained on the DHCP server. The address is assigned (leased) for a specific time period which can be hours or days depending on the server's configuration. If the device stays on line for the entire lease period, the lease is automatically renewed. If the device goes off line and then comes back on line after the end of the lease period, it may or may not get the same IP address.

Dynamic IP addresses should be OK for HMI devices such as RSView unless they are expected to receive messages initiated at some other device. The PLC that RSView is scanning needs a fixed IP address so that RSView can find it. The PLC doesn't need to know the IP address of the HMI unless it needs to send unsolicited messages.

 

[TConnolly]

Dynamic IP addresses should be OK for HMI devices such as RSView unless they are expected to receive messages initiated at some other device. The PLC that RSView is scanning needs a fixed IP address so that RSView can find it. The PLC doesn't need to know the IP address of the HMI unless it needs to send unsolicited messages.

Thanks Gerry. This is how we do it now. The HMI PCs and even the programming terminals get dynamic addresses and the PLCs get a fixed address. The problem is with assigning dynamic addresses to the PLC. I'm not sure its advisable, but we are running out of fixed IP addresses (we've only been given ten.)

I've actually thought about going to Linx gateway, putting two NIC cards in every PC, and putting all the PLCs on our own controller network where we have full control of it - I'd rather not take that step just yet - I do have to play nice in the IT sandbox. We are starting to venture into ethernet remote IO and I think I'll definitely have to do something like that there for IO networks - probably use two 1756-ENBT cards.

On a related note: I just got word that our local distributor is sponsoring a symposium next month where control engineers and IT spooks from the local area all get together with some Rockwell experts to talk control network integrity, security, and seeing eye to eye. Smart move on Rockwell's part. Really smart.

 

[Sheldn]

Alaric - I cannot enlighten the group on BootP and DHCP. I have used static IP. If you do messaging in CLGX using the old explicit MSG instruction, you need to define the IP address explicitly and the potential for the IP address to change will cause problems.

I have seen and it is usually recommended that you should use static IP for PLCs and HMI servers. Clients can use DHCP.

When you get together with Rockwell in your local symposium, ask Rockwell about what they plan on doing with "dirty network environments". It may provide some humor, or you may just get some blank stares.

 

[TConnolly]

I guess you better enlighten me first Sheldn...

What do you mean by dirty network environments? Do you mean dirty as in oil and dust? Or as in electrical noise? Or poorly designed networs? Or as in operators looking at girl pics on worktime?

 

[Operaghost]

Well,

BootP (which stands for Bootstrap Protocol) is nothing specific to Allen-Bradley/Rockwell. It is a standard networking term as is DHCP.

One of the advantages to using BootP is that your devices WILL receive the same IP address each and every time they power up, unlike DHCP. Plus the IP address can be easily assigned without the hassle of configuring.

For example, to talk to a new 5/05 out of the box, you must first create an offline file where you would configure the IP address. Then you must make a serial connection to download the project to the controller. Only then could you connect using the ENET port.

Using BootP the 5/05 MAC address and desired IP address would be entered into the BootP server software. Then when the unit is plugged onto the network it will be given it's IP address. You can communicate with it immediately.

Bear in mind though that this will not resolve your shortage of IP addresses. Your still going to be using the same number of addresses, you are merely changing HOW they are assigned.

Your idea of adding an additional NIC card or additional ENBT cards to create a "private" network is probably the better solution in the long run.

Also, be aware that if the BootP server is not available when the devices power up, they will not be assigned an IP address until the server returns. They will NOT retain their previous IP address. They MUST be able to access the server to be assigned an address.

BootP is very useful, but again it really isn't the solution to your problem.

OG

 

[Gerry]

Thanks Gerry. This is how we do it now. The HMI PCs and even the programming terminals get dynamic addresses and the PLCs get a fixed address. The problem is with assigning dynamic addresses to the PLC. I'm not sure its advisable, but we are running out of fixed IP addresses (we've only been given ten.)

I've never tried this, but there may be a possibility of using a DNS server to resolve IP addresses if the ENBT's have been configured with HOST names.

We are starting to venture into ethernet remote IO and I think I'll definitely have to do something like that there for IO networks - probably use two 1756-ENBT cards.

If you keep your I/O separate from your plant network on a dedicated ENBT and switch, then you can avoid the cost of managed switches.

That symposium should be interesting...

 

[CroCop]

If you keep your I/O separate from your plant network on a dedicated ENBT and switch, then you can avoid the cost of managed switches.

That symposium should be interesting...

After using a managed switch, and getting SNMP & network health reports set up, I'd never go back, the price isn't that bad.

 

[CroCop]

Alaric,

I got our IT guy to give us a class C network, with 2 NIC cards for my desktop. Full gig connection to my 24 port managed switch, and I couldn't be happier. We use the remote IO for almost everything except the overhead cranes, and the seperate network rocks.

Is the syp. in SLC perrty quick? Scott & I were thinking of going if we can sneak away for more then 5 minutes at a time.

Marc

 

[TConnolly]

Marc - get with Sharon about the symposium.

So do you use Linx gateway on the desktops?

I hooked up our PLC5s and SLC5/04s to our controllogix gateway today and viola, online over ethernet into DH+. I'm pretty pleased about that. I'm going to divide the DH+ network in two and use both channels of the DHRIO bridge and then kick the speed up to 117K. No more carting a laptop down to the floor just to make some requested tweak. Now I can retire two older ISA slot computers that I was nursing along to avoid buying PCI KTX cards. But thats what precipitated this, the gateway consumed another one of my fixed addresses.

 

[antoine]

Hello all,


Be very careful when using BOOTP option, during PLC startup, the ethernet port will send massages to the server to get an adress. These messages are broadcast and will go everywhere on the network.
Also, in case of server not responding it will keep sending these broadcast messages and then will overload the ethernet network and pull it down !

Hope this helps

Antoine

 

[Operaghost]

I am not a BootP expert, but I do use it a lot. A device that does not get a reply "should not" flood the network.

Most all devices that use BootP use a backoff method for sending additional BOOTPREQUEST messages. In other words each time you do not get a BOOTPREPLY, the interval increases before issuing another BOOTPREQUEST.

I do not KNOW for certain that the SLC uses this method (recommended in most tech manuals), however I certainly have seen this by monitoring requests received in my BootP server software.

Also, keep in mind that BootP messages are sent using UDP and will not pass through a router, also preventing "flooding".

In response to other comments made here regarding managed switches, it's not worth the few dollars you save to NOT choose one. Ok, so it's not a "few" dollars. Read up on the IGMP Snooping feature found on better Layer 2 managed switches and you will recognize the need for using one on an I/O network.

OG

** EDIT **

Some routers do have the capability to forward BootP requests. The feature is sometimes called BOOTP Forwarding or UDP Helper.

 

[CroCop]

Hello all,


Be very careful when using BOOTP option, during PLC startup, the ethernet port will send massages to the server to get an adress. These messages are broadcast and will go everywhere on the network.
Also, in case of server not responding it will keep sending these broadcast messages and then will overload the ethernet network and pull it down !

Hope this helps

Antoine

False. They have a setting to use last known IP addy if failure occurs.

Word to the wise, some program's BOOTP requests won't get past SP2 firewall unless you enable it.

Alaric,

I'm going to put my computer out in the new building when we get the proper permit to occupy, sometime this week (fingers crossed). I will use Linx gateway then. I had it setup on my laptop, but it was pointless as I don't have my laptop in the building everytime Scott does & vice versa.