Deploying containers to the Edge
As mentioned in the previous chapter, transferring your container image – what the Cloud industry typically refers to as deployment – is only part of the story when it comes to getting containerized microservices to the Edge. Recall the discussion in the Software and model deployment section of Chapter 3 stating the list of factors that can complicate the choice of where and when to place containers (and AI/ML models) at the Edge. To summarize, those factors include the following:
- The same software on different devices, servers, and locations
- Differences in CPU architecture on different devices and servers
- Specific co-processor and I/O dependencies
- Localized customization requirements
- Variations in operating schedules and procedures
- Differences in hardware generation, capacity, and topology
- Economic trade-offs between the far Edge and network or metro Edge in different geographies
- Differences in life...
