I'm tasked with implementing a single Edge Panel in a project to display data. The idea was to test out the functionality of an Edge Panel in an already existing project and the possible use of several Edge Panels in future projects. We currently aren't using Ignition but plan on migrating in the near future.
I've never used Edge products before and I'm not all that familiar with all the existing system architectures. I've seen the following architecture on IA's website;
This architecture displays what I've envisioned for the most part;
A standard Ignition Gateway is running on an iPC
A separate database connecting to the central Gateway
An Edge Panel connecting to the central Gateway
A device such as a PLC connecting to the Edge Panel
Is what I've envisioned possible with the aforementioned system architecture or should I be focusing on something else? Are there any other considerations I've failed to mention? What would be the minimal requirements for the iPC? The Edge Panel would be running Perspective which includes 2 sessions (1 for the Gateway and 1 for the Panel I presume).
If you are thinking of deploying both Edge and the central gateway in the same physical plant, don't. Just use a central plant gateway. (If you need redundancy, use a central redundant pair, not Edge fallback.) If you are concerned about network breakage in a single physical plant, upgrade to modern redundant networks (RSTP or similar) instead of using a fallback.
Don't use a conventional IPC for a central gateway. Ignition runs best with excess CPU cores and lots of RAM. Your database should be on a separate server in a production environment.
Your proposed architecture looks reasonable.
Some additional insight:
2 sessions are available for viewing the project running on Edge. The central gateway does not use up one of those sessions. You will have one available with your proposed. Though, it will be quickly apparent that clients should connect to the central gateway for viewing of projects. Plan to add unlimited client licenses there in the near future.
Plan to add EAM controller to full gateway. Manage all projects there, and push the project to Edge periodically.
If you need redundancy on your Edge device, contact your sales rep about Edge redundancy. Works wonderfully.
Don’t underestimate the difficulty required to configure a touch panel as an HMI on the Edge gateway.
Personal preferences:
Have all Edge devices establish an outgoing connection to the central gateway. Only incoming connections on central.
Never hand out the link to the Edge project. Use it for troubleshooting/testing among designers, or in the event of a second client on-premises, a Perspective Workstation client tailback (have second client connect to full gateway by default).
If adding several Edge gateways to the same central server, consider always bringing data in to the full gateway via GAN only (never add devices and polling engines, push that work load to distributed Edge devices).
I do plan on using redundancy, and in future projects we definitely will use a central redundant pair.
Currently I'm exploring options that Edge has to offer, one of which would be the Edge Panel HMI. For testing purposes, my idea was to have 1 Edge Panel connect to the central plant Gateway, and the SCADA would be displayed via the Edge Panel.
Does this make sense or should I look for an alternate solution? I'm required to integrate the Edge Panel in some way, and this option seems the most logical to me.
Network breakage is definitely a factor to consider. Seeing as I'm using several Moxa switches I most likely will use the RSTP configuration.
This is what I'm most interested in. For future projects we plan on having 2 separate servers - 1 for the central Gateway and 1 for the database.
Seeing as I'm trying to test out the functionality of the Edge Panel, my reasoning was that an iPC could host the central Gateway and a database well enough so as to not cause any major performance issues (again, this is just me trying to figure out the functionality of the Edge Panel on a small portion of an already completed project as well as the possibility of using Edge products on future projects).
For the iPC in question I was looking at the onLogic Helix 401. It seems to have decent enough specs and from what I can see, onLogic has a line of iPCs designed to be implemented with Ignition.
Thank you for your time, please let me know if I misunderstood something.
So from what I gathered, if I have the central gateway running on my iPC I can connect the session via the Edge Panel and a laptop, for example? My idea was to have 1 Edge Panel display the selected portions of my project.
For example, if I have 3 hallways (A, B, C) I would use 1 Edge Panel per hallway and display all tags that are needed in that specific area.
Would the central gateway be the EAM controller in this case or do I need to implement the EAM Module in some other fashion? My understanding was that the EAM was needed when there are multiple gateways in question - whereas I plan on using only 1 gateway.
As I mentioned previously, I definitely plan on using redundancy. Could you elaborate a bit on how you use redundancy with Edge solutions? Do you use a fallback or something else?
Got it, thank you for the advice.
Yes this is pretty much what I thought of from the beginning. I would have my Edge Panel have an outgoing connection to the central gateway and it would only be used as a tool to visualize data changes in real time.
Security is obviously important, I was planning on adding some sort of security measures just in case but I'll have to discuss that with the IT guys.
Well considering that we definitely plan on using several Edge gateways in the future this is something to be noted. When you say to bring data into the gateway via GAN, what exactly do you mean? Data from the PLC or something else?
Yes, when non-redundant. Consider having four when you set up redundancy. (Or two physical hypervisors with a GW & DB on each.)
If it is just for research, I understand. But putting Edge in the same plant as the Central GW it feeds data to is simply ridiculous. It adds such a load of code and project and data flow management that I classify that as "job security for SCADA staff". It isn't quite at the level of engineering malpractice as deploying any of this on Windows, but it is definitely bad for the end-user.
Sorry, @Chris_Bingham, but I think recommending this architecture in a single plant is really bad advice.
{ All of the above completely separate from the terrible choice to use Perspective as an on-machine HMI. Oy! }
I don’t believe I’ve done this. The post I replied to did not detail physical locations planned. While you and I differ on usefulness of Edge architecture, I think we both agree that it would be difficult to benefit from adding a second (non-redundant) gateway (Edge OR Full…) to existing on-site & flat architecture.
We will most likely go via the hypervisor route, but yes I was referring to the non-redundant part.
Could you elaborate a bit futher on this? I thought the whole point of Edge was to be able to add devices such as HMIs at the edge of your networks. Edge Panels have to be connected to a gateway in order to function, right? Is there any specific architecture that you would reccomend for my needs?
Maybe a dumb question on my part, but why is it a bad idea to use Perspective on an Edge Panel?
No, Edge IS an Ignition gateway, that optionally supports an HMI. "Edge of your networks" means geographically scattered.
If you want a machine HMI, just run a client on a touchscreen PLC physically mounted at the machine, connecting to your central gateway. (Preferably a redundant gateway.) With normal licensing, you can run hundreds of Vision clients from a central gateway. (Fewer Perspective clients.)
This topic is specific to machine HMIs:
This topic has more general comparisons:
This topic discusses considerations when coming from another SCADA platform:
This topic covers some server sizing considerations:
As I was reading through this, I think the general misunderstanding I was seeing (that maybe you've recognized now) is that Ignition Edge isn't required if all you need is a client on the plant floor. If you're going to have a central gateway running on something more powerful besides an iPC hopefully, you'll only install the client software on the local HMI computer (which works great for iPCs) that connects to the central gateway.
Edge is useful for cases where you're having remote sites or stations that are physically distant from the main network that need to run independently of the central gateway. It works well for standalone skids as a PanelView replacement. Where the skid is autonomous from other systems, and even then, if it's on the same network as the central gateway, you could (if you want) run the HMI normally off of the central gateway, but fail over to the local client in the event of a network break.
As Phil mentioned, if you're looking for a replacement for a PanelView, stick with Vision. Even on larger plants, I try to stick with Vision for HMI control of the plants, but when doing a SCADA system to monitor lots of plants and not primarily doing control from the system, Perspective works well for this. Use the right tool for the job. I've only once put Perspective on a local PanelView replacement where the customer wanted access to it from his phone on the local WiFi router in the panel and this was a standalone skid.
I would install Edge on an iPC and place a panel on the plant floor which would connect to a PLC for example. If that's the case, it's a much simpler solution that what I originally thought.
I'm not necessarily looking for a replacement, just some way to display the project. But you're probably right, I'll try to stick with vision for my HMI screens.
Yep, this is really what Edge is made for. Small local system with no database connectivity capabilities (you do have historical data, but it's stored in a built-in SQLite database that you can't directly query/access besides using built-in methods to query the historian). This is essentially a PanelView replacement on steroids (and cheaper too).
If you ever do get a central Ignition gateway, you can connect data from the edge to the central gateway through either the gateway network using remote tag providers, or if really distant (like over cellular modems), you can use MQTT to push data from the edge to the central server as MQTT is lighter weight than the gateway network in terms of bandwidth usage.
