ENTERPRISE ARCHITECTURE FOR THE INTERNET OF THINGS

enterprise architecture for the internet of things

By Mats Samuelsson

Abstract

Introduction of the Internet of Things in a company requires a completely new enterprise, business, and systems architectures. This article describes what IoT is, how it differs from Internet and current enterprise architectures and what new enterprise architectures (systems, technical, operational, and business) for IoT will look like.

It is an understatement to say that the introduction of the Internet required major changes in enterprise architectures. IT was suddenly not only managing internal applications but had to take on an external-facing web access function which grew over time from providing basic information to being fully integrated with enterprise functions like marketing, sales, support, logistics, production, documentation, and engineering. As organizations started to take advantage of evolving Internet capabilities, new functions and structures evolved over time. CIOs, webmasters, and SEOs suddenly became critical to the enterprise, and as web and mobile applications spread, so did the substantial operational IT headaches of keeping everything fully tested, functional, and operational.

The final addition is security, which is now taxing all web-based IT systems, forcing the establishment of special security architectures enforcing data security as well as operational integrity. The enterprise architecture for the Internet is still evolving with new approaches like micro-services introduced to address operations, scalability, and security.

With the Internet of Things (IoT), we leave the “simple” world where corporations, enterprise architectures, and organizations “only” have to worry about evolving web-based applications available through web browsers or mobile applications using HTML5 and APIs (application programming interfaces). Internet of Things suddenly introduces a new type of product to the business. This is the connected product that at regular or nonregular intervals communicates with new types of web servers—IoT servers—using local and Internet connectivity.

The concept of the Internet of Things evolved from the idea that with increasing processing power and connectivity (LAN, cellular, Wi-Fi, Bluetooth, ZigBee, 6LoWPAN) available at low cost (less than $10 and often less than $5) and packaged as complete “System on a Chip” (SOC), it suddenly became possible to connect millions of everyday products to the Internet providing new product capabilities that this connectivity allows.

Today we are starting to hear about “smart products”—smart cars, smart appliances, smart devices, smart toys, smart machines, smart transportation, smart cities, smart homes—in a new way, meaning that these products are connected products that interact with applications and services provided by the companies that provide the products.

Needless to say, this opens up a world of new opportunities for product companies to differentiate and add value to existing and new product. Internet of Things, with the promise it provides, is today where the Internet as we know it was in 1995. This also means that just as enterprise architectures were unprepared for the Internet then, the same can be said about the coming Internet of Things today.

Fig 1_Enterprise Architecture for the Internet of Things

Four Key Parts to IoT

So, what are these challenges? The Internet of Things introduces new systems and devices that have to be managed by IT and organizations not currently existing. The four key parts to IoT are: things (products), connectivity, IoT servers, and applications.

Foremost of these are the things—products that are the first key part of IoT—that now suddenly appear as something that the company has to manage.

At the beginning of the Internet, companies sometimes had to worry about browsers, but there was never a need to manage PCs. In more recent times, mobile app and phone compatibility required attention, although that is becoming easier with time.

In the Internet of Things world, a company now also has to take responsibility that its connected things (products) are really connected through the Internet to company servers (more about this later). This is kind of the equivalent to what a mobile operator does in making sure that all cell phones that it serves can actually connect to its mobile network. With IoT products connected through a variety of access methods, suddenly a new organizational responsibility arises within the enterprise architecture, that of making sure that hundreds, thousands, or millions of products are actually connected or can be reached by company IoT servers. With mobile operators, this connectivity is managed in what is called a Network Operations Center (NOC). In the Internet of Things world, each company needs to establish an IoT NOC managing connectivity—the second key part of IoT.

In the IoT world, products do not just connect with a web server speaking HTML5 or APIs. Each company decides which protocol and language will be used, and connectivity is to a highly specialized IoT server—the third key part of IoT. This server does not host applications but acts as an intermediary that provides four basic functions: device management and provisioning; data collection, processing, and normalization; application enablement and development support; and network data service and subscriber management.

In addition, many IoT servers provide functions like database and storage as well as security management, something we know is critical in today’s world. The world of IoT servers is today what the world of web servers were in 1995 before the standardization of two basic implementations (Apache and IIS). Today there are more than 25 companies offering IoT servers as enterprise software, cloud software, or as a service. Selecting the right IoT server is a challenging proposition since no market leader has emerged.

Furthermore, IoT servers require setup and ongoing operation, and most important, knowledge about how to program all core functions as well as setting up interfaces with IoT applications and how to interact. Differing from a web server, which is basically a flat file system with database interfaces, an IoT server is a rather complex system that requires trained personnel to manage and operate. That this system needs to be integrated with all other functions of IoT is a given, so enterprise architectures have to be created and modified to support this new function and capitalize on the business and operational opportunities it offers.

The fourth key part of IoT—applications—finally brings some good news into the picture in that it does not introduce any new components to the enterprise architecture but rather reuses the existing Internet application parts: web server, Internet and web browsers, and phone applications. However, this does not mean that existing organizations and architectures will suffice as they are. They have to be adapted so that they can serve the even broader set of external and internal applications developed as part of the new Internet of Things business opportunities and operations enterprise.

As pictured in figure 1, the Internet of Things will present substantial challenges to existing enterprise architectures and will require the development and creation of new IT systems and organizations as well as new evolving enterprise architectures. It would be a mistake to believe that existing IT and organizational architectures will be able to support IoT. Examples like IoT NOCs, new IoT servers, and a whole new IoT applications world requires the creation of specific IoT enterprise architectures.

Connected IoT products suddenly allow for in-depth understanding of how products are used, new ways of interacting with users and customers throughout the product life cycle, and the ability to collect and deliver massive amounts of product data and information. Capitalizing on this, enterprise architectures will evolve to support areas such as user engagement, product/service experience, product data/analytics, product customization, and life cycle service. Other areas affected will be R&D (now capitalizing on direct usage feedback), manufacturing (customized products), marketing (product/channel knowledge based), service (connectivity based), sales (usage-based sales pitch), and even finance (new revenue streams) and general management (better information). When we look back 10 years from now, IoT will have had an even more profound impact on enterprise architectures than the Internet did. A&G

Mats Samuelsson is the general manager of IoTgnosis, www.iotgnosis.com, a firm offering insights, advice, and training to companies getting started or implementing IoT in their businesses. He is a practitioner who has spent years with startups in Boston as well as Silicon Valley and has worked for corporations such as Motorola and AT&T. His specialty is to look at how organizational structures and skills have to be adjusted in order to support changing business conditions.

 

penn-state-online-ea-programs