Defining the Internet of Everything
Once Upon a Time There Was the Internet
A computer network is defined as a set of data communication systems interconnecting computer systems, and the Internet is defined as a network or a collection of interconnected networks. The Internet (with a capital I) is the largest internet in the world. The Internet is the global system of interconnected computer networks, linking devices worldwide and connecting public, private, academic, governmental, and corporative networks. A visualization of the Internet can be seen in Figure 1.1.
The technical foundations of the Internet leverage a core set of protocols designed with the guiding principle of maximizing interoperability. The Internet Protocol (IP) standardized at the Internet Engineering Task Force (IETF) is at the center of these core protocols and is responsible for end-to-end interoperable communications. At an infrastructure level, the Internet is comprised of hardware devices (e.g., routers, switches, access devices) and internetworking software. Layered over this infrastructure, we find a rich diverse set of services, which have made the Internet a valuable platform and, in turn, created a vibrant ecosystem. These services provide value by connecting computers that exchange information, and these new capabilities have created a deep social impact akin to an information revolution. The greatest impact of the Internet is not technological but social.
Figure 1.1 Visualization of the Internet
Source: The Opte Project by Barrett Lyon,http://www.opte.org/
Phases in the Evolution of the Internet
Looking at the evolution of the Internet, we can identify four very distinct phases or waves. Furthermore, each phase has a significantly more profound effect on business and society.
•In the first phase, the goal is to achieve networked connectivity, and consequently, digitizing the access to information.A couple of decades ago, just getting connected to the online world was miraculous enough. The World Wide Web (WWW) flourished, and the distance and time to access information dramatically shrank. People look for information online, send e-mails, and search the Web.
•The second phase is characterized as the networked economy, in which business processes are digitized. People buy products online. During the networked economy phase, e-commerce is developed, analytics finds a new anchor point, and the supply chain is digitally connected providing additional digital consolidation.
•In the third phase, networked emerging experiences, interactions are digitized. These digitized experiences belong to both the business and personal realms. This phase is dominated by digitally enabled social media and collaboration. Furthermore, widespread access to digital mobility compounds the effects of this phase where digital video for personal interactions is pervasive. People connect digitally with each other.
•The fourth phase, which we are entering, is referred to as the Internet of Everything (IoE). In this phase the whole world is digitized! We are entering a phase that unlocks an unprecedented value for businesses, governments, and society in general.
Figure 1.2 depicts these four waves and their growing impact on the world.
Each of these four phases builds on the previous one, and the technological and social advances of the past few decades have led to IoE. The IoE creates an incredibly unprecedented opportunity for connected services—an exciting new world.
Figure 1.2 Four phases in the evolution of the Internet
And Then There Is the Internet of Things
As we have mentioned, the growth of the Internet is occurring in accelerated waves. The acceleration is often explained as following Metcalf’s Law, or the network effect, which explains that the value of a network grows exponentially with the number of endpoints being connected (or proportionately to the square of the number of users). In other words, the more things and people connect, the larger the value of the network.
On depicting the size of the Internet as the number of devices connected, we can see that the milestones of 1,000 devices, one million devices, one billion devices, and 10 billion devices were reached, respectively, in the years 1984, 1992, 2008, and 2011. Figure 1.3 shows a logarithmic graph that helps visualize this exponential growth.
An even more significant milestone is that at some point near 2008, there were more things connected to the Internet than people in the world. Even more importantly, the slope of the growth is much steeper, with a rapid adoption rate of the digitally connected infrastructure. While the world population is growing linearly with a small slope, the Internet of Things (IoT) is growing exponentially, to soon reach a higher average of multiple Internet-connected devices per person. In other words, the adoption of the digital infrastructure is much faster—multiple orders of magnitude faster—than anything seen before by humankind, including telephony and electricity.
Figure 1.3 Growth in the number of devices in the Internet
The beginning of the fourth phase is about connecting the vast majority of objects (or things) that were previously unconnected. These things, as we will exemplify via use cases later in this book, are rich in variety and diversity. They include a robot in an industrial factory or a drone, a home meter, a traffic light, a pair of glasses, and a car. In the IoT, millions of new devices, sensors, actuators, and, in general, things are being regularly connected to the Internet. Furthermore, there is a shift in the endpoint landscape—from individual devices like computers, to things, smart objects,* and digital clusters such as the multitude of connected devices within a single car. Consequently, the IoT creates a network of networks, recursively.
From a couple of million things that were connected to the Internet in the year 2000, technology transitions and trends such as mobility, embedded computing, and cloud have started connecting the unconnected, such that at the writing of this book there were about 10 billion things connected to IoT. And still, most things are unconnected.
The IoT Reference Architecture
During the IoT World Forum Conference in October 2014, Cisco, in collaboration with Intel and IBM, announced an IoT Reference Model as shown in Figure 1.4.
The model breaks down the idea of data capture, data management, and data analysis into smaller subcomponents. It identifies the various technologies, their hardware and software components, how they relate to each other, and the boundaries and interfaces among the various layers. Understanding the interfaces and the boundaries facilitates a multivendor environment and ease of interoperability among the various layers.
This model describes interconnecting the things at the bottom layer (the edge). The things are any operational devices or elements: A temperature or pressure sensor in a factory, a security camera at a retail store, a wearable activity monitoring device, a cell phone, or any other device used by an internal process or by a customer or channel partner to conduct business or interact with others. The next layer provides connectivity among all these sensors or edge devices using a gateway or a connectivity layer that links local devices together. The model then introduces the concept of edge-computing.
Figure 1.4 IoT World Forum reference model
At the edge-computing layer, users acquire specific data from the edge devices based on a set of rules and then have the ability to conduct local analysis and produce results that can be translated into actions locally and without having to immediately transmit data upstream for further analysis. This does not imply that this is an analysis layer as much as it is a local action layer facilitating real-time action based on data in motion. If a particular process is approaching a threshold of some sort, this layer will give us an early alert before the data is processed at the upper analytics layers, at the same time it will send an action toward the lower layers to capture additional data to improve our sample or detect trends with a higher degree of accuracy (Figure 1.5).
The upper layers are concerned with managing the data from operations and processing and delivering it for analysis to the analytics entity for gleaning insights. The term process in this context is mainly concerned with the process data that has been integrated into business software like enterprise resource planning or customer relationship management to name a couple.
This discussion is a short introduction to the reference model. It will be helpful when we present, later in this book, detailed use cases about two industries implementing IoE capabilities to gain deeper insight into their current processes so that they can introduce data-driven changes to enhance the efficiency of their operations and improve their business outcomes.
Figure 1.5 Data-in-motion and data-at-rest
The Internet of Everything
So far, we have looked at the Internet through two different lenses: The Internet as an interconnection of devices and things and as a platform for human collaboration. While these two views are accurate, there is an additional ingredient of dynamism. There are two fundamental elements that complete the forward-looking story: First, people and things are interconnected to realize various processes. And in those workflows, there are significant amounts of data created. For example, the year 2012 created more information than the past 5,000 years combined. The Internet is ubiquitous.
These two additional items, process and data, complete the value picture of the Internet. Because of this, in May 2011, the United Nations called the Internet a “Fundamental Human Right.”†
IoE, which is the next phase of the Internet, is about connecting people, process, data, and things in new ways to unlock untapped value (Figure 1.6).
Figure 1.6 Internet of Everything
IoE, driving the next wave of dramatic growth of the Internet, will come from the confluence of people, process, data, and things. In a way, IoE can be thought of as a network of networks creating new connections that never existed before (e.g., a car, a traffic light, and a doctor) and spawning new opportunities as well as new risks. With increased processing power, decentralized architectures, context awareness, and real-time analytics, new services are constantly created. When these new capabilities are applied to the multidimensional platform, that is, IoE, they have the potential to dramatically increase global corporate profits, create new businesses, and invent brand new services. The IoE makes networked connections more relevant than ever.
IoE Connection Types
IoE connects both machines and people. This diversity creates the following IoE connection types:
•Machine-to-machine (M2M)
•Machine-to-person (M2P)
•Person-to-person (P2P)
The M2M communication, called the Internet of Things or the IoT, connects data sent and received to and from machines and things. In this context, the word machines not only includes computers, but also includes sensors, robots, actuators, drones, light bulbs, TVs, wind turbines, trains, mobile devices, and in general machines previously unconnected.
In the M2P communication, information is transferred from a machine to a person (or vice versa). This not only includes traditional Internet workflows such as downloading a web page, but also includes less traditional ones such as a human interacting with a remote point-of-sale machine at a kiosk. If a person retrieves information from a database or from a big data repository, something we call data and analytics also falls in this category.
Finally, P2P communication is often referred to as collaboration. Increasingly, P2P communication happens virtually (in instant messenger, voice, video, or a combination) as well as personally. These yet-to-be-invented collaboration means that connecting people together over the Internet is a key component of IoE.Figure 1.7 visually depicts these three connection types.
These connection types are not mutually exclusive. Most commonly, people will interact with things and other people on a single interactive process workflow. Furthermore, the business value increases when more connection types are used for a given interaction. Consequently, the highest value for IoE, where new business models, services, and innovation happen, is at the intersection of M2M, M2P, and P2P connection modes.
Figure 1.7 IoE connection types
IoT Versus IoE
One common source of confusion is the distinction between IoT and IoE.
In IoT, devices and things (i.e., sensors, robots, drones, etc.) are connected to the Internet and networked wirelessly or wired with standardized protocols, in particular the IP (see the section titled “The Internet Protocol”). IoE covers much more than merely things. IoE makes connections between things, people, data, and processes, and leverages analytics to make connections between apparently unrelated pieces of information. Information and analytics are the intelligence behind IoE.
In other words, IoT is a part of IoE; IoT is the subset that focuses on connections between machines. IoE, on the other hand, focuses on heterogenic connections among things, people, data, and processes (Figure 1.8).
The New Currency of Value
The IoE brings together four elements from the Internet as we know it today: people, process, data, and things. These four elements brought together create new networked-connections and unlock unprecedented value.
It is anticipated that the majority of the value created comes from new, seemingly unrelated connections being made of all the new elements connected to the Internet.
Figure 1.8 IoE versus IoT
One of the corollaries of adding connections to the Internet is the explosion in the amount of data. New people, things, and processes contribute massive amounts of data. This data, by way of analytics, gets turned into information (capturing the most useful pieces of data), which in turn gets correlated with other pieces of information to create insight. There is a clear distinction between the raw data, the captured and prioritized information, and the correlated insight. However, that is not the end of the IoE workflow. For this data to ultimately be of value, it needs to be turned into actions. Data and insight-driven actions create new capabilities and richer experiences (Figure 1.9).
Consequently, in IoE, the value (business monetizable value) is about sifting through immense amounts of data, realizing actions from prioritizing information, and correlating it into insights.
The main corollary is that what matters to IoE is the outcome of these connections. The connections themselves and the additional data generated, as well as the improved visibility, are a means to an end. The actions, business outcomes, better decisions, and new opportunities are the goals.
An Architecture for the IoE
IoE presents a completely new paradigm. As such, many of the commonly used architectural tenets and assumptions will not suffice or might not even be adaptable. At the core of this paradigm shift is the scaling element of an IoE. This section explores some key architectural building blocks supporting the IoE.
Figure 1.9 Turning data into action
Data
We have already seen a glimpse of the massive implications of the explosion of data as we connect previously unconnected things. Furthermore, data is the key ingredient to capitalize on the value of IoE, and data analytics is the engine to extract this value, it is paramount to understand what architectures can scale to leverage this data.
The requirements for data analytics of IoE are related to understanding big data. Connecting the unconnected brings new volumes of data. For example, 30 minutes of flight create 10 terabytes (10,000 gigabytes) of data by a jet engine. The world generates over two exabytes (over 2,000,000,000 gigabytes) of data daily.
And it is really more than just the amount of data. IoE brings diversity and variety in the sources of the data (e.g., data created by humans is structured differently than machine-generated or sensor-created data).
If we are to categorize the IoE data, we can find the following classes:
•Based on the structure of the data:
•Structured data, such as including clear semantics (e.g., XML)
•Unstructured data, without associated format and schema
•Based on the dynamisms of the data:
•Data at rest, for example, in a data center
•Data in motion, flowing as part of a business or industrial process
The speed or velocity of the generation of data as well as the need for results and actions calls for the creation of what is termed real-time analytics. Basically, the processing of data into action needs to take seconds, or even small fractions of second in some industries, instead of weeks.
Fog Computing
As we continue introducing new concepts and defining new terms, the one that follows is the most important ones in the context of IoE.
Traditional computing follows a centralized approach, in which data is moved to a data center, in a client-server model, to be processed. In an IoE world, the amount and velocity of the data does not allow for all data to be moved to a central location and then analytics be performed on it. For IoE, there is an intermediate layer between the end devices and the data center or cloud, and that layer is termed fog. The distributed IoT computing model can be seen in Figure 1.10.
Figure 1.10 IoT computing model
The term fog implies that it is like a cloud but closer to the edge. It is an expansion of the cloud paradigm, extending the architecture into the physical world. In the fog layer, some processing (e.g., analytics) is performed to provide actions close to the edge without having to send all the volumes of data to the cloud. The traditional data processing of Store→Analyze→Act→Notify is replaced by Analyze→Act→Notify→Store.
Fog computing was highlighted in The Wall Street Journal of May 19, 2014, as the “Tech’s Future.”‡
The final architectural building block is an interoperable protocol to provide connectivity, by way of location and identification. IP serves this role, and it can be seen in two versions coexisting in the Internet: the historical (legacy) IP version 4 (IPv4) and the current IP version 6 (IPv6).
Among many other things, IPv6 provides much enhanced and improved scalability, which is why it is the protocol of choice for IoT and IoE. For example, while IPv4 addresses are 32 bits, IPv6 addresses are 128 bits. While this book does not dive deep into the technical underpinnings of the IPs, curious readers can refer to the Internet request for comments 6272, “Internet Protocols for the Smart Grid.”