Billions of IoT devices exist, and this number will only grow as internet connectivity becomes a standard feature for a growing number of electronic devices. Although heavily integrated into the consumer electronics market, IoT extends far beyond handheld devices and home appliances. Industrial IoT aims to increase manufacturing efficiencies, and smart city initiatives look to transform urban environments.
Device management’s role in making sense out of chaos
In a nutshell, IoT device management is the process of authenticating, provisioning, configuring, maintaining, and monitoring IoT devices. Typically, this involves onboarding devices securely and maintaining the effectiveness of the device estate. This process consists of applying any necessary firmware and software patches or updates and keeping a watching brief on the IoT solution’s general health. IoT device management is continuous, throughout the device’s life cycle, from the cradle to the grave.
With billions of IoT devices already fielded and many more to come, the need for effective device management is only rising. Often, there is an automatic association of IoT – given the “Internet” piece – with the Cloud and an assumption that everything IoT must be done “in the Cloud”. While correct up to a point – after all, Internet connectivity is more-or-less fundamental to IoT – there are increasingly important factors that justify a rethink of this default mindset.
In this blog, the first of two, I will discuss the implementation considerations for an IoT device management service. We will then be exploring some of the nuances that might make the difference between an IoT solution that delivers benefit or one that adds a burden.
IoT + Device Management = Cloud…yes…no…maybe?
Leveraging a Cloud-based software-as-a-service product offering is undoubtedly an obvious option; it is a common approach and comes with various benefits. Indeed, the Pelion Device Management Cloud service is undoubtedly the right choice for many of our customers. After all, the barrier-to-entry is low, it can enable a quick-start solution, and scaling up (or down) is more-or-less painless. Accessing to the Cloud can be done from (almost) anywhere, and the flexibility that it provides, with today’s high-availability and geographic reach, is quite remarkable.
However, the on-premises type of deployment adds even more to that. It provides capabilities to customize the environment and integrate features and integrations that are not part of the default cloud offerings. It also offers a cost-effective and deterministic way to scale. Support for an expanded device rollout comes with low incremental costs; once established, customers do not pay for additional devices, transactions, or data storage.
Why can the right-sized on-premises solution provide significantly more value than using a public cloud service? Here are a few considerations:
- Host wherever you want. On-Premises enable flexible deployment on different targets: on physical bare-metal hardware, via virtualization layers (for example, Kubernetes), within corporate data centers, or even in alternative Cloud hosting environments. With an on-premises deployment, administrators can directly control every aspect of the implementation and access.
- Data resides where it should be. As organizations progress through their IoT journey, they find that the data derived from the solution is only increasing in value. Ingesting, processing, integrating, and sharing IoT data across the enterprise ecosystem helps to drive tactical and strategic business decisions. Increasingly, IoT data is also now a vital part of an Artificial Intelligence strategy. Given this rise in value, the IoT ecosystem cannot afford to lose control of their data. There is value in keeping the storage, processing, and management of this asset on-premises, logically close to their databases, servers, and applications. This approach also facilitates more straightforward integration with analytics, data-driven diagnostics, and other value-add applications.
- Regulations and laws favour on-premises deployments. Regulatory and compliance are among the top reasons that specific verticals cannot readily embrace the Cloud-centric approach. There are several reasons why these requirements might apply: legal or statutory, privacy, mitigation or indemnifications against threats, or market sector-based best practice. IoT is still not a standardized technology; there is no over-arching harmonizing industry standards body, which is an important consideration to bear in mind. Enforced regulations are not necessarily detrimental, and they often foster discipline, procedures, rules, and parameters that create value from their stakeholders. In technological terms, IoT solutions are not typically compatible with each other, and in the absence of industry standards, regulation helps to establish appropriate guard rails.
- Security remains top-of-mind for many seasoned IoT practitioners. The undeniable fact is that IoT devices are vulnerable to security threats. For an IoT ecosystem to be successful, a wide variety of technologies must come together. But this myriad of different embedded technologies also presents an increasingly larger attack surface. As we gain efficiencies and new benefits from the IoT, it would be a mistake to forget about the genuine security concerns associated with it. IoT brings with it high levels of privacy considerations. It is vital to secure the devices, the endpoints, the networks, and the transmission of data, to ensure the chain-of-trust extends end-to-end.
- Control of the System and Processes. Quite obviously, having direct oversight of the physical system enables better control over it. Relative to Cloud-based implementations, an on-premises deployment allows operators to provide instant access, administer system rights, halt the malicious attacks. Plus, prevent or mitigate many other problems that may arise over the life of the solution. Better control over the system also created transparency, provides visibility to all aspects of the system, and cumulates as more substantial control over other essential components of the IoT ecosystem: devices, connections, databases, and applications.
- New Business that only On-Premises can enable. Companies that manufacture IoT devices or provide IoT-based solutions operate in a very competitive marketplace. They are interested in building new business models, working with downstream customers, and innovating and delivering value-add services. One option is for customers to act as an aggregator of device management services, enabling a significantly different model from Cloud-based offerings. Leveraging an on-premise implementation, operators can easily configure modifications and system add-ons that further differentiate an innovative and evolving business model.
- Technical Limitations and Possibilities. For Cloud-based offerings, it is a one-size-fits-all approach. This aspect can be useful in terms of we are all in this thing together, and as new features are delivered, we all share equally in the evolution. Still, the downside is that – and this is undoubtedly true for IoT – most ecosystems are custom, and what works for some may not work for others. The other consideration is that incompatibilities caused by an unforeseen upgrade can wreak havoc with your SLAs. A relatively minor change in an API can cause connectivity or features to misbehave, causing the whole system to fail.
Where to from here?
While a Cloud-based product offering will be the right choice for some organizations, there is an alternative in the form of on-premises. If one or more of these considerations align with your IoT project, then an on-premises implementation may be the solution to your problem. Indeed, something of a best-of-both-worlds approach could be to deploy the on-premises implementation on a Cloud platform without open Internet access. This configuration delivers a more secure solution than the public alternatives, offers greater control and oversight than a generic Cloud offering, and has the sought-after Cloud agility.
Our latest On-Premises White Paper details the case for managing your IoT estate on-Premises and the criteria for choosing the right platform for you.