IoT Think Tank

Rise of the IoT Architect

Many organizations actually involved with the IoT have come to realize there's a need for a new role that brings other specialist roles together.

• By Clive Howard
• August 19, 2015

A recent CIC study of Internet of Things (IoT) implementations within enterprise organizations with systems engineering facilities addressing hardware and software solutions and products (for example, automotive, aerospace, medical equipment, manufacturing, electronics and telecommunication) surfaced some interesting insights. While many people speculate about the challenges and benefits of IoT, these organizations have actually been doing it. Although the definition of IoT may differ, each business was addressing the same fundamental issues: business case validation, data processing and management, networking, security and overall technology requirements, to name a few. As they progressed through these different areas, one thing became clear to all of them, and that was the need for new skills.

New Challenges Mean New Skills
While many of the issues faced by companies implementing full-stack IoT solutions (for example, those that include both hardware and software) resemble those that they have been facing for some time, it quickly becomes apparent that they are uniquely different. If we take the organizations with Machine-to-Machine (M2M) experience, for example, they have dealt with environments in which sensors are connected to central processing hubs in order to gather actionable data for many years. However these solutions have typically been highly proprietary from the hardware through to the networks to the software. IoT, by definition, introduces openness and interoperability into this world through open standards and open source. Whether that be in the use of standardized hardware, open networks (Internet) or open standards and protocols such as HTTP or REST.

So while many principles remain the same, the actual skills required and in many cases the processes also change. In the traditional M2M world, software updates were often minimal if not restricted to new hardware deployments. In IoT, software updates can be deployed frequently onto hardware that stays the same for long periods. In fact some companies see hardware lifespans increasing in an IoT world as users derive new value through software updates and don't have to purchase new devices. There are some clear examples of areas where skills will need to be acquired or bolstered, which I list below:

1. Network Handling:
   The Internet is perhaps the most disruptive element within IoT. Unlike the closed proprietary networks of M2M, the Internet is inherently open and that brings benefits and challenges. The benefit is that the network is simply there and does not have to be created. The challenges include that, in being open, the Internet presents threat agents (hackers) with a larger attack surface, and while it extends to every corner of the globe, factors such as latency and availability are less reliable. Many IoT devices will not plug into the Internet; instead they will connect via myriad of other networks including Wi-Fi, Bluetooth and Sigfox. Often these will be provided by different companies and so a typical IoT solution may be required to deal not just with different network types but different network providers.
   
   The connection may not be reliable but might drop in and out, meaning that in a request/response scenario one or other of the request/response may not complete. When the network is slow or not there at all, then the system must deal with that. All of this requires knowledge and skills that most organizations don't always have, although the need to handle intermittent connectivity has long been required for mobile apps. Once connected there can be a choice of standards over which data is transported from HTTP to CoAP. All of this has to be architected, coded and tested and potentially requires a number of partner relationships.


2. Security Management:
   Security is another area of equal complexity. Not only does the network need to be secured along with the data travelling over it, but also the hardware at one end and the processing hub at the other. Securing hardware can be complex in itself involving security at the device level (firmware). If you want to be able to access a device over the Internet, then you have to make sure that others cannot. Sometimes it's basic issues that need to be considered such as the fact that data travelling over the Internet takes many routes such that a message going from San Francisco to New York may pass through China. One has to be aware of this and ask if it is acceptable. People are clear that security needs to be baked into any solution and that will require new tools and technologies and the skills to utilize them.


3. Cloud Support:
   Many IoT solutions will involve cloud computing in some respect, and here again there are challenges. For many companies, especially large ones, cloud is still relatively new. For some others, it has not even been part of the conversation. It is also becoming increasingly complex to navigate to the right solution, especially given the cadence of new services and updates. Amazon is well known for an almost relentless stream of new capabilities. Some people find simply keeping up with this pace of change difficult. And then there are IoT-specific platforms and services that need to be evaluated against more generalist offerings. This is all aside from the decision of whether to utilize public, private, virtual private or hybrid cloud. Cloud often comes with new architectures and development skills, Platform-as-a-Service (PaaS) being perhaps the best example of this and a major paradigm shift for many. Much of this will be new to many organizations which will require the right people to identify and build out the right solutions.


4. Data and Analytical Processing:
   A key tenet of IoT is data and the insight it leads to, which is where the value actually resides. However, getting at that value will require analytics, another hot topic in the technology market right now. While there has been a proliferation of tools in this area recently and existing vendors have been seriously expanding their capabilities, they still require users with specific skills to get the most from them. The role of data scientist is often talked about and there is a shortage of such individuals. It is not enough either to simply have someone who understand analytics; they need to understand the system and the business as well. That means they need to be your data scientist not just someone contracted in.

A New Role Is Needed
Just these areas of network, security, cloud and data demonstrate the levels of complexity involved in an IoT solution and where organizations will need entirely new or perhaps more specialist skills. All of them need to collaborate in order to maximize the potential of any IoT project. What many organizations have come to realize is the need for a role that brings these others together. Someone who can have oversight across the many different disciplines (hardware, software, network, security, data and so on) and shape a complete solution. That role is often being referred to as the IoT architect.

Dawn of the IoT Architect
The architect position has typically combined domain knowledge, technical skills and the necessary authority to command various disciplines. Therefore it seems sensible to have just such a role dedicated to IoT with the necessary breadth in know-how to oversee the many constituent roles and teams. The challenge of course is that with IoT being so nascent there are very few people with the level of skills, knowledge and business authority let alone actual experience across so many disciplines.

Creating current architects can take years and some think that this role may take a decade to establish. Of course that means in the interim there will be challenges, but it also means that organizations who see IoT as being significant to their future need to start investing in such individuals now. The flip side to this is that there are current roles within organizations that already possess elements of the skills required by an IoT architect. As a result, it's likely that while there may be a dedicated IoT architect in long run, many will start by bringing together roles already harboring those skillsets into a central function or team.

Business Influence for the Technologist To Grab
Given the number of different skillsets needed to deliver IoT solutions, the broader impact on the business (for example, sales and support) and the potential to change or deliver new business models, the IoT architect role may need to command greater sway within the organization than other architects. That presents a great opportunity for technology functions within an organization to gain influence. But it also lays down a challenge to businesses who tend to shy away from empowering technology roles outside of their natural sphere. One thing that was clear from the CIC research study is that IoT will be transformational for businesses beyond technology. The result will be new skills, roles, processes, relationships and business models. Someone will need to have oversight across the solutions that deliver this change and that person could be the IoT architect.