In today’s data-driven world, a significant amount of data storage still happens on-premise. However, when organizations opt for cloud solutions, it is often through cloud subscriptions with hyperscalers like Microsoft Azure or Amazon Web Services. These cloud-based services offer tools for building applications, but the costs can quickly escalate, especially when logging thousands of tags per second.
For high-volume data platforms, the best choice often balances between on-premise or cloud server capacity.
As a founder of an industrial IoT software developer, I don’t advocate for on-premise storage nor for cloud services specifically. However, for those looking to implement IoT logging to improve their line of sight, it is important to distinguish between different services and their cost models.
One point that should be stressed is that while business models differ, most manufacturers have something in common when it comes to data logging: lots of data points. The sheer volume of data often becomes a concern in the long-run.
Business profile
The total cost of an IoT solution is a dealbreaker, especially when profit margins are tight. The pay-per-use model of cloud services directly scales with the amount of data that is
processed and stored using these services, leading to higher costs for high-volume data logging. In the meantime, on-premise platforms often offer a more manageable cost scaling factor.
The future of cloud storage for industrial IoT data is a complex question. Resources play a crucial role, as companies are hesitant to maintain their own data storage services indefinitely. As organizations scale up, they often consolidate their cloud services to reduce costs, potentially returning to more ‘traditional’ servers and software.
Cloud service business models are therefore effective for startups but can become less cost-efficient as companies grow – especially when IoT-data volumes grow along with them. This is a trend we’ve observed firsthand.
Cyber security
Security is another aspect worth considering when choosing an IoT solution. The belief that on-premise solutions are inherently more secure than the cloud is often an illusion. Modern technology and protocols for data transfer are secure, if not more so than typical on-premise systems that may lack regular updates.
When it comes to data sharing across borders, legal considerations may arise. However, in the case of production sites, this is – to most businesses and business models – not the most prominent concern. Major cloud service providers have policies in place to address cross-border data hosting, which often suffice for this type of data.
Critical connectivity
Connectivity, once a concern, has evolved significantly. Internet connectivity has become nearly ubiquitous, making unreliability a thing of the past. While certain control systems require direct on-premise connection for mission-critical purposes, the distinction between on-premise and online becomes less important for other systems. Many companies are therefore transitioning to cloud services for administrative tasks, SAP implementations, and visualized office environments.
Specialist solution
The main benefit of a specialized platform is that it makes for a ready-for-use, fit-for-purpose solution. However, as businesses scale their data acquisition, pricing models of IoT services can make it a pricey endeavor. In contrast, tailor-made software like ours is designed with these functionalities in mind, allowing for easy configuration and quick deployment.
It is a trade-off businesses have to make, based on their goals with this piece of software. The ‘build-versus-buy’ argument often comes into play. While you can use cloud services to build a platform with similar functionality as Factry Historian, for example, it will still require careful management.
In the end, the choice between on-premise and cloud services often results in a combination of the two, rather than a dilemma. Large companies may have their own data centers where some of their logging applications run, while satellite sites operate in the cloud. Central instances allowing multiple production sites to connect to a single cloud setup are frequently favored, making hybrid setups quite common.
Future scalability
We advocate in favor of future planning. Whether you start on-premise or with an appropriately priced cloud storage solution, business should plan for scale. Businesses that keep in mind the thousands of data-points that will be logged in your system, realize that a pay-per-use model for the logging itself is often not the best route. It simply becomes too expensive in a manufacturing setting, where volumes are massive.
Therefore, it’s wise to plan before you buy – to plan for future scalability. Considerations such as consolidation of cloud services or the potential for transitioning from on-premise data storage to the cloud should be factored in during the design phase. Don’t hesitate to seek guidance from experienced professionals, and learn from examples in your industry. Doing it right the first time, can save you a lot of time, money and trouble in the long run.
About the author
Jeroen Coussement is the CEO of Factry. Headquartered in Ghent, Belgium, Factry helps process industry customers create clarity using IIoT solutions that digitize production activities and increase effectiveness. The scale-up brings its solutions and open data culture to companies—from SMEs to multinationals—on five continents and in 27 countries. Visit Factry.io to learn more.
