The distributed hybrid topology extends the supply chain management enterprise cloud hub, with one or more vertical scale units in the cloud or on Edge, running business processes in vertical domains near or where the work occurs. It helps to increase resilience to provide higher reliability, extra performance, and maximum uptime.
The main values of the distributed hybrid topology for supply chain management are as follows:
- Operations scalability with workloads allows applying new workloads for warehouse and manufacturing processes as you scale your business without affecting system performance. Likewise, you can move workloads between deployments to optimize your supply chain operations;
- Resiliency against connectivity issues – No impact on warehouse and manufacturing processes when the connection to the Cloud hub is lost, or the connection has latency issues due to geographical distances;
- Availability across maintenance windows – Hub and scale unit deployments can be serviced at different times and run different versions. For instance, it allows warehouse and manufacturing operations to run uninterrupted on-scale units during maintenance windows on the Cloud hub and vice versa.
From a deployment perspective, there are four main options at the time being. The most likely to be used is an enterprise cloud hub – traditional Dynamics 365 Finance and Operations deployment, whereas all system components are deployed in one or more instances in a Microsoft Azure-hosted environment with Microsoft to be responsible for administration and maintenance of these environments on your behalf.
You may also want to deploy with new recently introduced options: Operational Cloud Scale unit and operational Edge Scale Unit. The cloud-scale unit is deployed in a separate instance within an Azure data center, be it the same or a different one. Still, it will be hosted on the Azure Cloud with administration and maintenance is performed by Microsoft on the scale unit.
With the Edge scale unit, you would deploy server components at the customer site, either at a facility, warehouse, or manufacturing facility, to run the supply chain management solution. Your server components would be installed locally at the site and would be managed and maintained by the customer IT department.
Also, there is one more way to deploy, called Tactical Edge. Those teams who deploy finance and supply chain management main application components via the local business data option (LBD) may want to choose this one. Tactical Edge deploys on-premise components locally at the customer site, meaning both finance and supply chain management main application components and main additional scale units installed at the customer locations.
The next question would be how scale units end up working at all. And the answer would be workloads themselves.
There are two primary workloads available at the time being: WES (warehouse execution workloads) and MES (manufacturing execution workloads) workloads.
For instance, you’d be using your warehouse mobile app to transit against those workloads with WES. Moreover, you’d have a production floor execution interface transacting against those MES workloads, whereas transactions would be getting queued up on those scale units and then, either synchronously or asynchronously, depending on connectivity, would be synchronizing those transaction updates over to that main hub.
From the hardware perspective, there are specific Azure regions for F&O traditional deployment in an Azure center hosted by Microsoft.
Within data centers, edge zones are available for scale units at various data centers around the globe for teams that choose to deploy these scale units.
And so, either within that same data center as your hub or a global footprint across data centers, you can deploy these scale units in the cloud within Microsoft data centers around the globe.
If you choose to move forward with any of the edge scale units or the tactical edge scenarios where you are deploying the hardware locally on the customer site, make sure to choose recommended hardware to be installed at those locations.
Finally, keep in mind that this runs on top of the local business data infrastructure for the edge scale units, so it’s not quite as simple as just dropping in a single server and these locations.
Essentially, when the synchronization occurs, the scale units take ownership of that data from the hub. All “update,” “create,” and “delete” actions need to be run against that scale unit while it owns the data, which will then get synchronized back over to the hub once those data updates are completed.
This is how to prevent duplicate data or data issues from occurring in distributed hybrid topology.
From the licensing perspective, skews are now available for both cloud and edge scale units. This is an “above and beyond” standard supply chain management licensing to deploy scale units, making it a per scale unit licensing model.
There are two options available, and each supports a different number of transactions. But, you can add additional capacity depending on your needs.
Hence, MS has a basic and standard cloud-scale unit add-ins. The basic one covers 200.000 base transactions per tenant/month, with an additional 10.000 capacity that can be added by purchasing these overage licenses.
The standard is a 1.5 million transactions per tenant/month capacity package with an additional 100.000 transactions capacity that can be purchased.
Let’s recap. Working within the cloud hub or cloud and/or edge scale units requires a loosely coupled architecture for the chosen business processes. Furthermore, it will be synchronous to asynchronous communication channels, depending on which deployment option you choose.
There is a concept of scale units and workloads – so the scale units are physical environments hosting the workloads (so those are the servers either in the cloud or on-premise). Then we have workloads that are defined and allocated to these scale units.
One thing to note: since you can mix and match, depending on whether it is, let’s say, a warehouse facility or manufacturing facility – if there is a facility that does both, you can deploy multiple workloads to those scale units, so you don’t need multiple scale units, or it’s not one for one relationship – you can have multiple workloads on a single scale unit.