IoT and networking infrastructure

Internet of Things (IoT) can be described as connecting any device with an on and off switch to the Internet (and/or to each other). This includes everything from cellphones, coffee makers, washing machines, headphones, lamps, wearable devices and almost anything else you can think of, including jet turbines, cars and oil rigs. IoT is where objects will come alive, pay attention, interact with each other and with us, our friends, our co-workers, customers and the people we meet every day.

 The number of connected devices is expected to exceed 25 billion by 2021 according to Gartner. These connected objects will monitor, understand and anticipate our needs and above all respond to our choices. Many of us already use IoT, on a regular basis. From wearable health monitoring devices, smart surveillance systems to predictive maintenance and supply chain optimization in the industry. IoT is pioneering the next phase in the Internet’s evolution.

 As individuals, the IoT revolution will totally transform our lives, we will have fewer constraints and complications, greater peace of mind and more time to connect to what matters, resulting in improved quality of life. When it comes to business, IoT promises to make organizations more agile, more efficient, improve competitiveness, create new revenue streams and new ways to serve your customers. It enables companies to increase their efficiency and innovation capabilities. All this and more, thanks to the way IoT will optimize processes and create new solution offerings, as well as reinvent customer relationships. IoT will bring cities and communities into the "smart city" era: promising them a greater understanding of their citizens’ habits and more efficient management of public services. New traffic management and waste collection solutions are already coming into force, along with initiatives to optimize energy consumption, protect our planet and improve quality of life for the citizen.

 These are the reasons why many companies have started realizing the importance of IoT and started investing in connected technology. According to Statista, more than 80% of senior executives across industries, on average, said IoT is critical to some or all lines of their business in 2018. For further IoT business statistics, see Forbes article: 2018 Roundup Of Internet Of Things Forecasts And Market Estimates.

Where to start?

There are many ways to begin the IoT journey. Don′t let the complexity of IoT scare you from exploring and participating.

 After doing the GTM plan, a simple way to commence the IoT journey is to lay a robust framework for your IoT system. There is no single architecture that will suit all areas of IoT. The key is to understand what suits your plan best, and consider future needs. Today, many enterprises invest in a four-part IoT network architecture to support IoT systems. Treat the devices as stages in an IoT process. All of them are integrated structures that transport valuable data from different networked ‘things’ to production and IT systems to deliver actionable business insights and take actions.

1.  Sensors and actuators. In this stage, sensors collect data from the environment and turn that data into useful information. Data processing takes place in every stage of IoT architecture. The data that is processed at the sensor has limited use, as the processing power available on each IoT device is limited. Since data is at the center of IoT network architecture concept, one must also consider immediacy and depth of insight while processing the data. The faster the information needs to be delivered, the closer to the end devices the processing must be executed.

2.  Internet gateway. The data that is collected in the sensors is in analog form. Hence, it needs to be curated and converted into digital streams for further processing. The sensor network is connected, collects outputs, and converts the data. The internet gateway receives the collected and digitized data and sends it to stage 3 systems.

3. Edge. The digitized, data can now communicate with IT systems. However, the data may still require some processing before it can enter the cloud or data center. This is where edge systems come in. Edge systems perform more analysis of the data. They can be placed in remote offices or edge locations but are usually located in the facility closer to the sensors.

4. The Data Center and the cloud. Data that requires more in-depth processing gets forwarded to cloud-based systems or a physical data center. At this point, more efficient IT systems analyze, manage, and store the data. It is already a long process where data must reach the fourth stage. Commonly, you can undergo an even more in-depth analysis or combine sensor data with data from other sources for deeper insights. The processing in this stage takes place in cloud systems, in a hybrid cloud system or on premises.

But then, how is everything connected?

As more enterprises evolve their IoT proof-of-concept projects into live architectures, IoT won’t be the only technology migration many are tackling, as enterprise networks are also evolving into a broader, more multifaceted transformation. IoT is coming into play just as enterprises are transforming their IT infrastructure and migrating beyond the hub-and-spoke architectures that have defined networks for decades. In a hub-and-spoke model, all services are processed in a centralized location, and all connectivity goes through that hub. All enterprise traffic from that hub might get backhauled through one or more private links (eg MPLS). It is a model that reflects a previous hardware-centric enterprise IT era, and do not offer flexibility to prioritize applications or traffic, to access applications from different locations and device types, or to host and process applications in one or more external clouds. There are several converging trends in recent years that require a new network approach: the growth and variety of different devices - not just enterprise desktops, but smartphones, IoT sensors and other devices - connected to the network; the proliferation of more distributed networks and remote work; the ever-present need to reduce enterprise costs; the rise of new network connectivity technology options, like broadband Internet access and 4G LTE (and now in roll out - 5G); and growth in hosted applications in a variety of places, not just on a laptop or a data center, but in a diversity of potential cloud locations.  Enterprise networks are growing more complex. IoT devices being added don’t have the same network usage rights as other devices. The explosion of applications adds complexity. Applications are everywhere, in the data center, in the cloud, or in multiple clouds. On top of that, complexity of BYOD in corporate environments. At the same time, the user expectations are the same performance whether they are in the corporate enterprise, in a remote branch office, at home or in the café. Branch offices are directly accessing the Internet to get the right application experience, so the security need is now distributed in nature. The hub and spoke environment anymore don′t apply anymore.

 Forget the old M2M model with one very simple device in the field connecting to a cloud. IoT will require a very different type of architecture where it just won’t be feasible to send everything back to a central cloud. Instead, it’s about pushing a multi-tiered architecture, generally referred to as edge. There are many types of edge, ranging from someone’s factory to a data center. These represent different tiers in which various levels of processing are occurring. Gartner refers to this model as Mesh Application Services Architecture (MASA), and it’s all about varying where you do compute depending on where it is most advantageous to do it. One use case is to process workloads at edge in order to save bandwidth costs. Another example would be in finance, where even a millisecond of lag in a trading algorithm computation can mean a substantial financial return. Edge computing architecture allows us to place servers in data centers near stock exchanges around the world to run resource-intensive algorithms as close to the source of data as possible. This provides them with the most accurate and up to date information to keep their business moving.

 The issue from the traditional network point of view is that the way of networking historically has been managed is incompatible with these massively distributed architectures and multiple tiers of compute and storage. Operating proprietary hardware, circuits and VPNs – the old way of coping - doesn’t offer you the agility and flexibility to connect to anywhere or scale as you need. The next generation WAN is the Internet, imagine the reach! That’s where we need to focus on building MASA if we’re going to drive the business outcomes that enterprises want. This is a challenge, requiring cloud-native connectivity that solves all these problems. 

Different data centers providers have different strategies for the problem. Many are responding by providing more for the cloud service providers – more power, more space, more locations for hyperscale cloud to reside. Others are going in the opposite path, all about pushing more to the edge. The major telecommunication carriers are also thinking in that direction because they’ve got lots of legacy locations where their core network goes. If you have a lot of distributed edge locations, they may consist of only a few racks. There may not physically be space to connect multiple circuits into it, as you would do if it were a major data center where many networks collocate. So many are opting to go with one proprietary option, which is a potentially dangerous strategy.

 A few years ago, a new concept was introduced, software-defined wide area networking (SD-WAN) with the aim to simplify all of this complexity and help evolve enterprise networks into more flexible, programmable architectures that can meet the changing expectations of users.  SD-WAN accomplishes this by adding software overlay to the enterprise network that separates network control and management functions from the physical network, like what software-defined networking can do in a data center or public carrier network. SD-WAN is a network edge architecture comprised of both overlay (software-based) and underlay (circuit-based) components. SD-WAN may have an important role to play in enterprise and industrial networks but many IoT applications do not yet require large amounts of bandwidth on short notice. SD-WAN-based visibility into multiple enterprise connections and control of entire enterprise capacity pools will allow dynamic allocation of bandwidth for mission-critical IoT applications as they emerge, while also segmenting the most latency-sensitive and security-sensitive applications of the industrial IoT. Typically, you see SD-WAN in front of or part of the IoT gateway or a highly optimized service residing on the microcontroller architecture. SD-WAN’s ability to rapidly scale for increased traffic loads will become a significant advantage with the anticipated growth in connected devices for a certain set of use cases. Nevertheless, companies such as Netflix and Walmart can deploy code thousands of times per day. How can you accurately update your network configurations and firewalls that frequently? How can you instantly provision private circuits and SD-WAN CPE from IoT to cloud, and everything in between? This is the mismatch. We are moving towards agile, cloud native, DevOps automated continuous development and deployment, while the network still relies on manual configurations, siloed, separated groups and slow provisioning processes. There are concerns about SD-WAN security problems, agility, higher than expected costs and complexity?

 We need to run our apps everywhere – across Internet, edges, clouds, service meshes, mobile user and IoT devices - in a programmable, automated manner. We need to enable developers and apps to programmatically define the networking and security they need – to enable your apps and DevOps tools to programmatically control any network. We need Connectivity-as-Code. Developers, architects, DevOps and DevSecOps can programmatically define what the network needs to deliver, rather than be constrained by what has pre-provisioned on a network, or rely on the network itself to provide security, reliability or performance. Organizations require centralized visibility and control of all application connections, regardless of what WANs or clouds they traverse, while the edge innovators in those organizations are free to use their existing DevOps, application and service mesh orchestration tools to meet their connectivity needs. Control should be managed by software, rather than the handcuffs of proprietary networking, manual configuration, private circuits and separated teams.

Application Specific Networking is an existing disruptive and innovative solution, that capture best of both worlds and addresses these needs and challenges. It uses the public Internet with software-defined network overlays. Imagine enable users to “spin-up” virtual private networks on-demand and with the transformative economics of cloud computing. It turns the ordinary public internet into a zero trust, secure, performant and cloud native enterprise class network by enhancing it with next generation zero-trust cyber security, while at the same time boosting "best effort" Internet resilience & performance. The job then is to build highly secure, application-specific connectivity, all the way from the IoT device, overlaid on top of any of the tiers of architecture that are doing the processing, right back into the cloud. One secure overlay secures everything, from the data in motion to the end points. In this way you are ensuring minimal risk exposure by implementing security-by-design with things like zero trust, micro-segmentation of applications, trust-based connectivity from hardware certificates and software-defined perimeters. At the same time, you will exploit public Internet′s existing assets and get global reach while ensuring performance-by-design and bring extra performance and reliability to the connectivity. It is all software, based on DevOps principles, and can be orchestrated on-demand with cloud-native tools.

 Companies buying into this new software-defined approach to IoT are seeing much less complexity and much more agility. They are providing an easy environment for IoT devices to connect to. Plus, developers of IoT solutions are enabled to focus on what they do best – app development - innovating, creating new use cases, with all the complexity abstracted away that you’d have if you decided to build your own connectivity into a data center. 

 The IoT revolution is still in its infancy. As the IoT expands and becomes more widely adopted, it will provide new opportunities to ensure meaningful progress for people, society, and the planet. Gone are the days when Software-Defined-Networking (SDN) was called “Still Done Nothing”. Embrace the software led business revolution.

Stefan Johansson