When it comes to fundamental IoT deployment strategies, problems are often pitched in starkly binary terms. Do you host onsite or the cloud?
But like most things that initially seem to have no common ground, the real answers are more nuanced. Not only is the right answer a little different for every business, but there’s no need to go all-in on one solution or the other; hybrid strategies often ring the bell.
Understanding the basic differences
As industrial IoTs become more common, IT managers grappling with the array of sensors, instruments, data pools, and other elements increasingly need to decide how to manage the applications, data, and administration that comprise these networks.
The distinction between cloud and on-premises is often easy to see. Self-hosting on the premises means you have total control over all the components. You can choose where and how to store your data. Your data may be more secure, since it is stored locally and you can exert your own security protocol (though that might not be the case – in many cases, cloud solutions can offer better security than on-premises). And using on-site storage means you can more easily customize your solution and readily integrate it with in-house tools or third-party applications.
The cloud, in contrast, is more about simplicity and ease of use than customization and control. The hosting is done for you; you’re given simplified setup tools, but can’t customize or integrate anything the cloud provider doesn’t enable for you. But support is standing by — generally 24/7 — and the cloud provider is the one guaranteeing uptime, which conceivably means you can save money by not staffing to guarantee mission critical coverage.
Understanding the difference in cost
Costs are a critical factor as well. Most cloud services charge on a pay-as-you-go basis, so you’re only paying for services and storage you actually use. Keep in mind that if you decide to self-host, you need to make extensive investments in hardware to accommodate your IoT needs.
And then there are ongoing costs to maintain servers and storage space, pay for power and maintenance, and more. Further ongoing investments will be needed to keep your systems current and in operational. When it comes to the cloud, your provider shoulders all those expenses.
All that adds up to a strong case for hosting in the cloud. But, especially in the industrial IoT, it’s not a black and white issue.
You need to consider issues like: How much data do your IoT devices generate, how quickly do you need to process it, and do the various instruments, sensors, and equipment have inexpensive and reliable access to the cloud. Not every extremity of the IoT has hardwired internet access, for example. How will you get data to the cloud? If you need to lean on a cellular connection, for example, then data costs can add up quickly. Nor will data transfers necessarily be fast.
Moreover, a lot of data coming from the IoT setup can be redundant, so you might want need to filter out unneeded data before transmission. All that suggests a hybrid hosting approach.
The benefits of a hybrid approach
In a hybrid approach, you combine local and cloud-based hosting. In this scenario, for example, you may choose to both store and process data at the edge of the network. You might analyze data, reduce and filter, and then move it on to the cloud. Or you may choose to keep certain data on premises permanently.
You can opt do this because you need rapid access to the data and don’t want to need to pipe it though the cloud to process and analyze it. There may also be security concerns; especially sensitive data you may opt to keep out of the cloud and host locally, either in centralize data lakes or at the edge, where it’s processed.
Regardless of the exact reasons, it’s clear that there’s room for a mix of on-premises and cloud hosting in any IoT project.
Steve Latham is the founder and CEO of Banyan Hills Technologies, an Internet of Things software company located in Duluth, Ga.