We are looking at IoT deployments for our manufacturing plants. I’m confused about where "Edge" ends and "Hybrid Cloud" begins. Is the edge just a mini private cloud, or should I be managing it through my public cloud provider's edge services (like AWS Outposts)? How do we handle data synchronization when the edge sites have intermittent connectivity?
3 answers
Think of Edge as the "fringe" of your Hybrid Cloud. The goal of Edge is to process data as close to the source as possible to reduce latency—crucial for real-time manufacturing sensors. Using services like AWS Outposts or Azure Stack Edge allows you to use the same APIs you use in the public cloud, but on hardware physically located in your factory. For intermittent connectivity, you need an "Event-Driven" architecture. Your edge devices should store data locally and use a "Store and Forward" mechanism. Once the connection is restored, the edge gateway syncs the summarized data back to your central data lake for long-term analysis.
How are you planning to handle the physical security of these edge devices in a factory environment compared to a locked-down enterprise data center?
Edge is essentially just decentralized hybrid cloud. It’s about putting the compute power where the action is, rather than hauling every single raw byte back to a central region.
Well said, Nancy. The "data gravity" at the edge is too high to ignore; processing locally saves a fortune on bandwidth costs alone.
Joseph, that’s a major concern. To answer your question, we are looking into hardware with "TPM" chips and full-disk encryption. Since these boxes aren't in a traditional data center, we have to assume someone could physically access the device, so the data must be encrypted at rest and during the sync.