- June 14, 2018 at 2:35 am #2346
I have a question regarding the use of Unmanaged vs. Managed switches in Ethernet/IP environments, specifically regarding where exactly along the line is it OK to use either. I know that Ethernet/IP uses Multicast Traffic, and if you use unmanaged switches they treat Multicast traffic like Broadcast traffic.
A managed Switch will use IGMP Snooping/IGMP Querier to determine where to send that traffic, and remember what ports it should dent it too. This way you dont flood your network with a bunch of requests, and cause latency, Jitter or even PLC timeouts.
I also know that PLC’s consume multicast traffic, and I/O produces Multicast traffic. So if you have a lot of I/O and its not isolated then you can run the risk of these kind of issues, but if its isolated than its usually OK.
Any advice on where and when Unmanaged is OK to use with Ethernet/IP networks, i want to advise some clients on why they shouldnt use Unmanaged switches to add more ports to their Managed Switches.
DGJune 14, 2018 at 8:27 am #2347
Hi David, great question! This is often a debate and it looks like you’ve got a good handle on when a managed switch should be considered. I would add a couple of important features you get with managed switches that you won’t get with unmananged:
- Port Mirroring (for troubleshooting traffic)
- Port Diagnostics
- SNMP (for switch management)
Like you said, things like I/O, motion and particularly HMI traffic can push a lot of traffic on your network – especially if you are going with cheaper alternatives (non-Rockwell).
That said, you can consider unmanaged switches under the following conditions:
- In isolated networks – networks are segregated from each other and not interconnected via a router or other means. The need for multicast traffic mitigation is greatly reduced.
- If you are using only Rockwell products – Rockwell products are well designed and can tolerate high load multicast traffic where IGMP has not been employed.
- Multicast capability for every node must be greater than the sum of all multicast traffic in the network. For example, the 1756-ENBT module can tolerate 5000 packets per second. This is a sum of multicast from joined and non-joined group members.
Of course, without doing traffic analysis (via a managed switch) knowing what kind of multicast you have is next to impossible. I hope this has been somewhat helpful and that other will chime in.June 16, 2018 at 9:55 am #2447
I just wanted to take a minute and thank you for taking time to answer my questions. I also wanted to give you the recognition you, and your team deserves for creating not only the content that you do…
But also the community that you continue to build, Love your Site and Love your Content!
Your answer is both concise and very informative, I sincerely appreciate it.
Have a great day Bud!
DGJune 16, 2018 at 10:11 am #2449
Thank you for your kind words – it really does mean a lot! What I’m trying to do here is build a modern, clean and robust platform for us all to share our expertise. Thank you for being a member I really do appreciate it…enjoy your weekend and if you ever get inspired to share a project, topic or idea via our blog platform, I’d love to hear from you… just visit our Guest Post page to get started!
FredJune 16, 2018 at 10:30 am #2450
I actually do, I am interested in understanding how to approach the factory automation process.
In doing my research your site is absolutely doing just that, Building a Modern clean easy to navigate place to learn and share.
I have found so many outdated difficult to use repositories of automation information, usually, they are full of threads that would be useful but?… There was nobody there to moderate and help make the conversation beneficial, they are usually all over the place.
You know what i would love to see but have not found, so I am working on creating?
an application recommendation diagram that shows you in a very easy to understand way, what type of PLC and what type of Network switch to use and why?
With Graphics, and model names and Technology Terminology like IGMP Querier/Snooping, in Multicast traffic situations.
I am looking into the different type of Rockwell PLC’s for instance, and I can’t find anything that says use this one for this type of application. or even Real Life uses of Factory Talk to monitor not just the switch status, but also how sensors are being utilized more and more to make troubleshooting easier for plant floor operators.
Because they can use additional Sensors to give Abnormal/Normal/Fault status alerts and also walk them through on what steps to take in either case.
anyway, keep it up my brother, you’re doing the Lords Work!
DGJune 16, 2018 at 11:06 am #2451
Hey David…that sounds very interesting. I certainly think this type of “selection workflow” type information would be useful. Almost like a decision tree of sorts…if you have this many IO…these types of devices…connecting into this type network…eventually spitting you out to a leaf or series of leaf node type options with hardware specifications…This is what I’m understanding from your description anyhow, sounds great and rather ambitious.
When you’re ready share what you’ve got, and perhaps we can collaborate back and forth, send me an email [email protected]June 16, 2018 at 11:17 am #2452
I mean honestly, I know that there is no one size fits all model but having the Lowdown on whats what?
would make it so easy to figure out by using your Application requirements, because they would be used as a reference for selecting the model etc…
at the very least if I can help someone understand the Managed Vs. Unmanaged, how I/O helps with simplifying troubleshooting and therefore makes it easier to train any less experienced operators. Therefore helping alleviate the stress of hiring from a less than an ideal pool of candidates.
That’s what makes automation such an interesting topic, all the unforeseen benefits that most people don’t always consider.
I’ll definitely take you up on the offer, your advice thus far has been more valuable than you can even imagine.