The Internet of Things (IoT) continues its explosive, exponential growth as the “killer app” for both the cloud and for edge computing. And it’s easy to see why from a number of angles:
- For manufacturers and product owners, IoT enables new, connected features that increase capabilities, customer “stickiness” and, ultimately, revenue growth.
- For technology vendors, IoT drives the potential for substantially increasing the consumption of services, while strengthening customer engagement.
- For technology buyers and users, IoT provides many benefits, ranging from better visibility, new features and functions, increased utilization, reduced errors, improved safety and decreased costs.
The appeal of this is immediately apparent: the more connected devices a business has, the more opportunities to improve visibility and make better decisions on how to operate.
Large enterprises, especially those that are distributed, with many locations and assets to monitor, are struggling with the surging need to connect, maintain and effectively manage all the devices. At the same time, as enterprises begin to realize the disruptive potential of IoT, they are struggling with how to rapidly meet the changing needs of the business while building solutions and infrastructure that will scale securely and cost-effectively.
The Lean Startup Philosophy
Four years ago, when my last startup was incubating a new automated IoT logistics solution, we decided to build it as a cloud-based service running on AWS. At the time, this was an obvious choice. We were firm believers in the Lean Startup, a popular book by Eric Ries, in which “failing fast,” “minimum viable product (MVP),” and “DevOps principles” were the order of the day. Building on top of AWS quickly provided us with global reach at reasonable cost while we developed the initial MVP, secured our initial customers and scaled the company. We leveraged a variety of the then-available AWS services, such as compute, storage and database, but by and large we had to develop most of the core services ourselves using open source technologies and our own homegrown analytical models.
Fast forward a few years. If we were starting today, our design and architectural approach would be very different. In the fast-moving world of startups — and of today’s most innovative enterprises as well — time-to-market is absolutely critical to any initiative’s chance of success. In the case of a startup, it also can mean the very survival of the company.
Therefore, today there would be a low-to-zero percent chance that we would build a platform from the ground up. Rather, we would develop our solution on top of someone else’s cloud-based IoT platform. But which one would we choose? As we say in the consulting world: “It depends…” Indeed, the answer would very much depend on which industries, solutions, geographies and markets our products were seeking to serve.
We often hear concerns from our clients regarding “cloud-lock,” where clients seek to “maintain optionality,” and then make design decisions that rely heavily on generic, open source products and tools. However, they make this compromise of perceived code portability at the expense of leveraging native, highly optimized, cloud vendor-provided services and features that can dramatically increase solution performance.
The exact same argument can be made against “platform-lock,” where choosing the wrong vendor may limit future integration or deployment capabilities. While there may, in fact, be a trade-off in terms of absolute, long-term flexibility when making a platform decision, in many cases the benefits gained in capability, consistency and time-to-market far outweigh the risks of lingering platform-lock concerns.
While every client has its own set of technical and business challenges, as consultants we often observe patterns across clients as we deliver solutions. Our clients value the experience, frameworks and reusable assets we bring to the table, which in turn allow us to implement best practices quickly and achieve the shortest possible time-to-value.
One of the most effective ways we achieve this is by focusing on, and building competency in, what we believe to be the best-of-breed IoT platforms that are quick to deploy from an infrastructure perspective, and that subsequently allow us to build and deploy robust business applications and solutions on top.
In other words, there are a lot of factors tied up in helping our clients choose the right platform for their needs, and ensuring that it is quickly adopted by the business, its employees, partners and customers. Choosing the right platform can unlock tremendous value and set our clients on the path to success. Choosing the wrong one may lead to project delays and missed opportunities.
The Secret Sauce
So how do we help our clients choose? The secret sauce for IoT platform selection has as its key ingredients and considerations.
First and foremost, there is no off-the-shelf, one-size-fits-all IoT platform. All such platforms are works-in-progress, as are the cloud platforms on which they run. It is safe to say that they will all continue to add new features and functions, hopefully MVP-style, and on a predictable timeline on which we can help our clients plan their own development activities.
Many of our clients have asked us to design a “foundation IoT architecture and platform” that will work across all the complex and evolving needs of the business, to support both current known and future unknown needs. This is an admirable goal and, in fact, there are indeed many common components across IoT platforms that address fundamental capabilities, such as device management, provisioning, security, data ingestion, analytics and event processing, and communications.
However, even after choosing the “right” platform, unique client-specific environments and challenges can make implementing many of these features far more complex than clients initially realize. Sometimes the urgent desire for speed in building a solution can come at the expense of solving for such common needs.
The converse can also be the case, where clients deliberately seek to address all current and future needs of the business (long-term security, trust, privacy, large-scale ingestion and data storage, large-scale device management, etc.), but at the expense of time-to-market. Driving this approach there is usually an admirable goal of “getting it right the first time”, but designing an architecture for its own sake is seldom tied to real customer value. This can lead to a variety of other adoption issues, not to mention friction between IT, OT and the various lines of business demanding applications. The best IoT platforms seek to address many of these challenges, while creating an environment on which new business applications can be quickly built and deployed.
We help our clients through these critical considerations to come up with the right secret sauce for IoT platform selection. Here’s a quick list of the things we address when evaluating and choosing an IoT platform that will meet their current and future infrastructure, development and business needs:
- Market focus: Is the IoT platform vendor focused on connected consumer, commercial or industrial products? Do they have a laser or shotgun perspective?
- Hardware compatibility: Will the platform integrate with and speak the languages of the client’s existing OT assets, platforms, devices, networks and systems?
- Software and technical skills: Does the client’s team have experience with the programming languages required to build apps on this platform? Are apps readily available in the market, or are they scarce?
- Ecosystem: Can and will the client’s partners and customers use the same platform? Will their hardware and software vendors support it?
- Security model: Is there flexibility to support different use cases?
- Deployment model: Does the platform support multiple models and provide global coverage (i.e., edge, on-premise, public cloud, multi-cloud, etc.)?
- Autonomy and control: Does it support disconnected/remote use cases?
- Analytics, machine learning: What are the out-of-the-box capabilities to power the client’s algorithms and speed up their development time?
- User interface: Can the client control the branding and/or build engaging experiences that will increase adoption by employees, partners and customers?
- Flexibility and openness: Will the IoT platform support multi-vendor devices, networks, multiple protocols, languages, edge and core services, etc.?
- Integration capabilities: Will it easily integrate with legacy and enterprise systems?
- Future-proof: Is the solution extensible, easy to evolve for new IoT use cases and technologies?
- Support model: Does the vendor provide adequate coverage (language, hours, geography, etc.)?
- Pricing model: Does the vendor’s model reasonably fit the needs of the client’s business?
- Vendor viability and risk management: Will the provider be around to deliver support in the years to come? Does the client have an appetite for risk? Is the platform software in widespread use (which would tend to have defects discovered and fixed)?
By now it should be fairly obvious that there are no perfect fits, and any attempts to analyze an IoT platform vendor, RFP-style, in a matrix, may leave the evaluation team scratching their heads and frustrated.
However, it’s very likely there will be one or possibly two IoT platforms that will be the closest fits to the unique needs of the client’s business. Choosing the right one will set them on the path for meeting the needs of their employees, partners and customers for many years to come.
CTP’s IoT expertise and approach can help you get to your goals quickly and stay focused on time-to-value solutions that grow your business.