Much of the technology needed for the Internet of Things (IoT) has been available for some time, but connecting a world’s worth of devices to the cloud requires more than just an Internet connection. For IoT deployments to be successful, business models and pain points for players on both the device side and network side must be identified, followed by solutions that can collect and analyze data securely and efficiently.
The IoT; intelligent systems; surround computing. Whatever you choose to call it, the big buzz in embedded over the last couple of years has been around ubiquitous computing and how connecting edge devices to the cloud will change our perceptions of data. However, once you get beyond the buzzwords, the potential, and even the technology, the IoT is nothing more than a concept unless it addresses the business models and use cases of the enterprise and carrier networks.
In the context of the IoT, both carriers and the enterprise face significant deployment challenges, with businesses having to consider how to integrate new and existing devices into a “system of systems” approach, and network operators needing to respond to the data traffic produced by billions of sensors. Both must also determine how to move data northbound to the cloud for analysis and then back southbound to execute on gained intelligence, and how to manage it along the way (Figure 1). Until recently, these setbacks were the result of a lack of technology. But with the advent of Software-Defined Networking (SDN), Network Intelligence (NI), and manageability tools, the discussion turns to how companies can incorporate these technologies to use the IoT to their advantage.
“The technology ingredients are there today [for the IoT], but not necessarily the business models and supply chains,” says Gareth Noyes, Chief Strategy Officer, Wind River. “We are going to need to deploy infrastructure at multiple levels – whether it is at the edge, gateway, sensor hubs, or the cloud – that is scalable, both in terms of performance, price, and other things. One of the reasons why the embedded space is so fragmented is that we are dealing with price points, different business models, different technology demands. That leads to us having to be very scalable in that approach.”
“You have to identify where the pain points are on the carrier side and the enterprise side,” says Jim Douglas, Chief Marketing Officer, Wind River. “The starting point is, ‘what value is the IoT going to create?’ The IoT is a big concept, but if you break it down in its simplest form, it is all about trying to unlock the value of your assets and business processes; trying to make them perform better; trying to make sure they are more sustainable; trying to ensure there is no disruption.”
The business case for connecting “everything”
As mentioned, the business case for the IoT is that it allows carriers and the enterprise to make better use of existing infrastructure. For carrier networks, the IoT presents the opportunity to provide tiered services for data that is more or less critical. This can be achieved through software solutions that enable new revenue streams and also limit additional overhead.
“From the carrier side, the question is not how will the networks be able to handle the IoT, but how the carriers will get paid,” says Douglas. “How do they quickly drive down Operating Expenditures (OPEX) but make money at the same time? It is an issue of effectively accelerating information through data pipes, which is a lot more effective with data acceleration engines. Content inspection also allows you to look for bad actors, as well as implement some flow analysis and traffic shaping. This enables you to see what is going through the pipes, then manage, analyze, and monetize it.”
“If you do not have enough physical bandwidth in your pipe, or the data volume is too high on your device, you are going to need some intelligence to throttle that in an intelligent way,” Noyes explains. “Unless you own all of the infrastructure, you do not have control over the latency and data paths. Today you are leveraging either public infrastructure or leased infrastructure. That does not necessarily give you the underlining performance that you need. That is probably going to be a business opportunity for service providers – to provide different classes of infrastructure. You may pay more for something that has low latency or high determinism, and there are services that have adapted to the ‘good-enough’ model – video streaming as an example. On the other hand, in a lot of end-user applications, where you have maybe security or safety-critical applications, ‘good-enough’ is not good enough.
“The concept here is if you are providing different types of traffic over a single IP pipe – if you are transmitting video, monitoring data, tweets, e-mails, and a bunch of things all over one pipe, you are probably going to want to have a better quality of service with protocol data,” Noyes continues. “What the Intelligent Network Platform allows you to do is look at different flows and on a packet-by-packet basis, determine what you want to do with the packet (Figure 2). You can make it high priority or more critical.”
Intelligent networking enables the IoT business case for service providers and, when combined with SDN technology, helps facilitate the northbound/southbound flow of information. On the device side, the IoT will provide enterprise businesses with actionable intelligence about their resources, allowing data to be extracted from the edge, analyzed, and returned. However, executing on this process requires that intelligence is applied at the bottom layers of the cloud so that information can be retrieved from disparate devices in a secure and digestible fashion. As many existing devices were deployed without regard for the IoT, gateway architectures can be layered in to preclude extensive field upgrades on legacy systems.
“We have devices everywhere, and a lot of these were purpose built to be safe and secure,” Douglas says. “The way they accomplished that was by being isolated from things like enterprise networks where nasty things happen. A lot of high-reliability systems in Aerospace and Defense and carrier networks have taken the network out of the question, and many apps have tried to stay off the Internet. This results in no leverageable data. This data can be used for predictive maintenance for remote repairs, to get information from a lot of devices to reduce OPEX, and for adaptive analytics so you can adjust elements to make everything more optimum.”
“You cannot have a closed, proprietary, non-API for this infrastructure,” Noyes explains. “It is going to be based on interoperable building blocks, standards, compatibility, well-defined structures, and APIs. We have a platform called the Intelligent Device Platform that is architected specifically around an open operative environment that caters for different management protocols, whether it is OM, OSGI, OMDA, and other protocols on the management side (Figure 3). On the communications side, there are things like ZigBee, Bluetooth, and other things. We are already architecting programs like the Intelligent Device Platform that have the building blocks, and we can fully expect to incorporate others. That is on the edge side facing down to the devices.”
Different strokes for different verticals
Flexible solutions are a critical component for IoT deployments due to fundamental differences in industry verticals. Though the technology and business cases are available today, there is still work to be done to optimize the IoT for individual markets. This will have to be conducted on an industry basis, and requires ecosystem collaboration to unlock value in the core network and at the edge.
“Ultimately, you love technologies that can go horizontal because that makes some really big plays,” says Douglas. “But in the interim, you have a bunch of vertical markets that have unique situations right now – they have different standards for connectivity, different APIs, different protocols for different peripherals – so you have to attack it on a vertical basis. The other thing is that the value chains are slightly different in all of these industries, and so how you actually deliver that value is going to be different. We are going to have to partner on a vertical basis with ecosystem players that are trying to serve those individual markets.”