The journey of Mordor was never for the incautious. The all-powerful and omniscient Eye of Sauron could wreak havoc in split seconds, and fast decisions were to be made at that granularity. That is why Frodo Baggins, instead of going all alone, was accompanied by Samwise Gamgee, who stood up against all odds to make their journey worthwhile. If you’re wondering why an article about the Internet of Things lays out the plot for J. R. R. Tolkien’s Lord of the Rings, have faith because it will be worth it. Amidst his complex yet fantastic universe, Tolkien offered us a pearl of essential wisdom – Two is not just company, two is a team.
Internet of Things has just begun unfolding its life-changing features. The smart speakers and the wifi-operated home devices are just the introductions to the disruptive changes that IoT would bring for the digital age. We’re talking complex surgeries, sensitive gas pipelines, air traffic control, and much, much more. Sooner or later, industries would be readily accepting IoT for its unparalleled offerings. This is possibly why the IDC report suggests that by 2025 an average connected person would be practically engaged in one interaction with an IoT device every 18 seconds! Therefore, the IoT data, which is most crucial for such far-reaching implications, cannot be left to be minded by just a singular storage and data management infrastructure. We need a Sam for when Frodo can no longer operate coherently.
Fellowship for the ‘Things’
Based on the presently available scope for the Internet of Things, we have two main concerns regarding Data Management.
- Latency – As the performance granularity gets down to milliseconds, we would need as low latency as possible. The slightest delay in the decision, and we might find ourselves dealing with a global crisis.
- Durability – While scalable performance is indispensable, we would also need a durable infrastructure that could store the massive amounts of data (we’re easily talking zettabytes here)
Cloud Storage is a powerful storage infrastructure for scalable data processing and persistence. However, the intense data processing situations that the IoT networks would be working with may cause network congestions while interacting with the clouds. This may lead to unaffordable delays, not to mention the risk for Stuxnet-level cyberattacks. On the other hand, the NVMe-based edge computing can easily interact with the IoT sensors and processors and provide them the required latency, but it would be a nightmare if chosen for ultimate data persistence. Alone, either of the architectures would ridiculously fail to serve the glorious purpose of IoT. However, together these two can provide the best ecosystem for Industrial Revolution 4.0.
The Symbiotic Architecture
The collection, management, and response of the data in real-time are all that IoT is about. In addition to that, we would also need a powerful architecture to make system-wide actions as and when needed. These are the grounds that project the necessity for the Cloud-Edge interdependence we talked about earlier.
Cloud Storage, with its global connectivity and peer-to-peer architecture, can easily deploy high-volume data objects at the edge locations. Similarly, the NVMe-based edge devices can collect real-time actionable data from the IoT network and dump them on the cloud at regular intervals. We will now breakdown the utility of the system
The actionable data will be generated by devices like sensors that would essentially observe the behaviors like air traffic or pressure in gas pipelines. Such data cannot be trusted with the cloud because of network security and latency risks. Therefore the NVMe-edge layer would provide the necessary data collection hub with the required speed and scalability needed to make real-time decisions. With advantages like multipath I/O, multi-stream writes, and asynchronous events capture, NVMe serves as the best option for edge data collection and processing with just the right precision and efficiency. This architecture can also serve as temporary storage for the collected data before it is dumped at a more permanent location.
Even if the cloud storage cannot provide the extreme processing speed and precision required by the IoT devices, it will still be needed to store the humongous data volumes that would be collected at the edge. Moreover, object storage is known for its scalability, durability, and availability. This means that once the data object are stored on the cloud, they can be computed for more long-term analysis and behaviors that would help the industries predict and avert macro-level threats and inconsistencies. The cloud architecture promises higher protection standards for the data owing to its backup automation and data recovery advantages.
Finally, the analyzed data needs to be visualized and presented in structures that would help the organizations make important decisions for ongoing and future projects. For this, we would need a dashboard both at the edge as well as cloud networks. At the edge, the dashboards would ensure the autonomy of the IoT devices by taking appropriate actions in real-time. While at the cloud level, we can trust the dashboard to help us with system-wide alerts, important notifications, distributed data structures, and global data visibility.
As per a report, the global IoT market is almost certain to reach higher than US$13 billion by 2026. Even during the pandemic, we saw the inefficiency of healthcare infrastructure that can only be handled with technologies like Edge computing. Therefore, we need to start working towards an end-to-end architecture for incorporating the Internet of Things at the industrial level. This article aims at providing the basis for such architecture. One where different storage resources can come together to serve a greater purpose. Cloud storage and NVMe-based storage can be the Sam and Frodo we need for this arduous journey. Alone, they might not be enough, but together they can certainly defeat even the most powerful threats.