Will the Internet of Things revolutionize cloud?
If your car needs data, it is not looking for it in the data center. If your windmill needs data, it is not coming from the cloud. In both cases, and in all cases, when we look at the Internet of Things (IoT), the data is generated locally and is to be used locally, at the device.
In addition, the decisions that are made based on the data, need to be done fast. Not only in the case of cars (“is that a child there on the road”), but also because our companies and our consumers are asking for quicker answers and reactions; drive by the increasingly digital and 24/7-around-the-world economy.
This latency, the time needed to get the information from the data center, is driving a new 3-tier architecture in IoT. With the sensor, the data generation, at one end, the Edge-device in the middle and the data center, physically or in the cloud, at the other end.
This Edge-device can exist in 3 types; a simple device that collects data from multiple sensors, a more complex device capable of running simple applications and a third variant that resembles a mini-data center that is capable of more compute and storage power. For sake of simplicity we can call them a Monitor, a Gateway and an Edge server.
The need for this compute and storage proximity to the data will result in a situation that Gartner recently predicted:
“Around 10% of enterprise-generated data is created and processed outside a traditional centralized data center or cloud. By 2025, Gartner predicts this figure will reach 75%” (source here)
This growth in the importance of the Edge computing layer in a 3 tier IoT architecture, creates some unique challenges in areas of security and systems management.
From a security point of view the increase in Edge devices creates a much larger surface of attack, both physically as digital. And from a systems management point of view the distributed nature of the Edge computing layer creates unique challenges:
- More physical devices to manage, without centralization of labor
- Deployment of workloads across a vastly distributed compute platform
- Managing resources on devices with limited compute and storage capabilities
- The provisioning of the devices itself; physical distribution and up to date software resources.
But there are benefits as well; lower latency being the first one that comes to mind. In addition, we can foresee that this 3-tier architecture will grow into a more mesh like structure when Edge devices will be able to communicate directly with each other. Such a cluster of Edge devices can strengthen the individual compute and storage capacity through bundling of resources.
Does this all mean the data center becomes obsolete? Absolutely not in the near future. Especially in cloud based data centers we can much more easily scale up compute and storage and allow for the creation of complex analytical models and machine learning; the trained models could then be distributed to the Edge devices for local utilization. This will result in the ‘smart edge device’ that can act autonomously, possibly even off-line for certain periods, while it will connect regularly with its neighbors or its parents in the data center to see ‘what’s new’.
At the Atos Technology Days 2019, Paul will explore how IoT can help businesses gain the #DigitalEdge. Find out more about the event here.