Cloud Computing

[AgentSmith]

Have you implemented MES in the cloud?
Have you experienced any gotchas?
What provider did you choose and why?
What type of connectivity are you running?

[DeterredNerd]

We have also been considering cloud computing for digital manufacturing. I see tons of people talking about AWS and Moderna, I really wish someone from that project would weigh in on some of the questions that you asked here. Thanks for this wonderful website!

[COOL]

I have not. In my view cloud has three benefits:

1. Removing undifferentiated heavy lifting (a bit of a spectrum depending on the level of cloud technology used (Iaas or Paas))
2. Pay as you go pricing
3. Increased Speed to set up and get going from scratch.

Based on my experience with MES, the only salient benefits are Pay as you go pricing and speed of set up. I don't know all the MES systems, but assuming their like my prior experience, you still have to configure those servers and install software, as they are essentially "lifted and shifted" to the cloud.

Am curious if anyone knows an MES thats truly structured as a service? It would remove concerns related to networking, servers, OS, software and simply expose the APIs that add business value. A customer could start your MES in minutes. It would take away that undifferentiated heavy lifting.

[house_stark]

Thank you for starting this topic. Korber / Werum PAS-X is available as a service: Körber Pharma | Werum PAS-X as a Service -Cloud MES software & intelligence products (koerber-pharma.com)

But I don't have any experience with it. Would love to hear from others that do, or experience with any MES as a service (assuming there are others?). As an MES engineer, my concerns (maybe unwarranted?) with using a MES SaaS system are around loss of flexibility. Will the SaaS provider allow you to connect directly to the DB to query if you need to? If your business process requires customizations, can you have these in SaaS? But if your requirements are fully met by the chosen MES functionality then maybe these are not important. Also I would worry about other companies (Microsoft? Amazon?) having control over shutting down manufacturing in my company if there is an outage. I'm sure there would be SLAs in place, but guess I feel better knowing infrastructure colleagues at my own company will be working hard to restart the MES servers to allow manufacturing to continue instead of some random Azure technician who could care less about manufacturing running or not. But this is probably an old-school way of thinking and I need to modernize. Am very curious to hear others thoughts on this.

Thank you again.

[COOL]

It’s a real concern. The classic way to manage outage risk of course is redundancy, in particular, cloud providers have availability zones, physically separated from each other in different geographical regions in case of service outages. That way if something goes down in a specific region resources in another kicks on. Also the cloud providers tend to roll out service updates region by region to ensure minimal disruption.


This is a benefit of the cloud at least if you’re using the higher abstraction services (blob stores like s3 or compute like server less functions) tend to be built with redundancy baked in, no work required on your part. It’s more work if you have classic applications running on cloud instance, since you need to configure a load balancer. Also I guess there’s networking that could go wrong, redundancy again here can help, setting up a dedicated connection to the cloud is commonly just in case internet doesn’t work.


Definitely something to consider though.