Other edge computing requirements

- [Instructor] So, what are some of the other things we should consider to be successful with edge computing? Well, keep in mind that security is still a key point. And you need to secure edge systemically. In other words, from the sensors to the device, that's gathering the information, your edge computer, to the back-end cloud-based systems. So you have to deal with encryption, you have to deal with different security parameters, and that needs to be built into pretty much everything that you're dealing with with edge computing. Next would be operations. Keep in mind, once we roll something out, that if we're successful the first day, it needs to be successful the thousandth day. Therefore your ability to have consistent operations, and your ability to maintain these systems, whether it's fixing edge computers and allowing automation to automatically reroute network issues, the ability to, in essence, take the complexity that is in edge computing system and operate it in such a way we're automating as much as we can, and we're providing access or visibility into the various systems of the edge-based computers using abstraction or the ability to simplify the views, and simplify the management. Governance means we're putting guard rails around the edge-based system. So for instance, perhaps not allowing people to change the database within an edge based device, unless they leverage some governance systems to make sure that other people are notified that the database is being changed, or the ability to put guard rails around use of services or use of AI systems. This is really keeping you out of trouble. Now this is not security per se. Even though it is limiting access in some instances, in some instances it does work with security. It's really the ability to protect ourselves from ourselves. So in other words, we're not able to get in and do things that we didn't mean to do that actually hurt the system, because this protections in place. Management is very important. And it's really co-located and co-exists with operations. So not only do we need to operate the thing long-term, and have typical operational playbooks in hand like management monitoring the ability to deal with performance systems, the ability to deal with security updates, all these sorts of things that really kind of go into the business of running any system including an edge computing system, but the ability to manage it longer term. That's updating hardware, software, dealing with configurations, things like that. Data management is very important considering the fact that we're storing data in many different places. Sometimes at the edge device, sometimes in an edge computer between the edge device and the centralized cloud-based system and always within the cloud-based system. So if we're dealing with information, we're typically dealing with information via edge computing via tiers. So therefore the information in some cases is it redundant between the device and the back end cloud-based system. But in many instances, it's unique. And therefore the only copy of that data is either maintained in the cloud or maintained on the edge based device. We do this because we're trying to save space. So if you store the same information in two different places, obviously you're wasting storage. However, it makes data management that much more important because keep in mind that we have to protect the data, we have to deal with the schemas, we have to deal with the master data management, we have to deal with the backups and the updates, all these sorts of things need to occur as well. Other concepts to consider include the fact that we're integrating artificially intelligent and machine learning based systems. Now we're doing so at each tier. In other words, we're not just leveraging machine learning on the back end cloud-based systems, but we have intelligence at the edge. Therefore machine learning is being leveraged at the edge based device as well. Therefore it's able to make learning decisions and make instantaneous decisions as to how to act upon data that's coming from the sensors. The ability to manage hardware. You got to remember, we're dealing with devices here. In many instances, we could have tens of thousands of devices that are out on a factory floor, in a farmer's field., any number of applications or use cases can be there. So therefore we have to have the ability to update these systems, update the operating systems, we have to go out and repair them, we have to replace the memory cards, we have to upgrade them over time, we have to deal with networking issues, and that means traditional hardware management. The ability to deal with network management as well, keep in mind, this is the lifeblood of the edge computing system, 'cause the edge computing system communicates with the devices via the network and communicates with the backend system as well. And then finally, the device-to-cloud management, your ability to manage the chain between the sensors and the back end cloud, all the devices and computers that sit between them and do so in a holistic way and in doing so we're able to keep the system healthy.