Featured Topic

Embedded Computers Optimized for Cloud Deployments

After fifteen years of growing anticipation, the Internet of Things (IoT) is at last taking material form, heralding a new age of machine-to-machine communications. At this year’s Mobile World Congress in Barcelona, IoT and M2M (machine-to-machine technology) were the words on everyone’s lips, with the first hodgepodge of early solutions on display everywhere: home monitoring systems integrating consumer devices with residential automation systems, medical machinery for remote communication with hospitals—a wide array of consumer and industrial appliances looking to create new, automated uses and habits with smarter cars and wearable devices, all ready to communicate with one another. IoT has arrived, with Gartner estimating that by 2020 it will be a 300 billion dollar market.

If the Internet of Things is already big news among retail manufacturers, it’s even bigger news in heavy industries like energy and power, or shipping and transportation. For companies delivering industrial commodities the possibility of rolling out cheap, massively distributed networks that can turn consumer and process data into hard profits is a tantalizing and clearly understood opportunity. The only challenge is finding computing platforms that can do the work. IoT is the natural endpoint where industrial automation completes its convergence with information technology, and with a full line of integrated hardware and software solutions built specifically for industrial cloud applications, Moxa is ready for it.

The Trouble with Massively Distributed Deployments

One of the first challenges for system integrators and other suppliers of massively distributed systems is how to securely and reliably connect edge devices to the central core. In this context, security and reliability mean something far beyond narrow IT-versus-IA considerations.

On the one hand, there is the sheer scale of the deployments: without strongly secure, effective automation of monitoring, maintenance, and management processes, integrating thousands or tens-of-thousands of edge devices into an effective, centralized control system is a herculean task, and one that may well be too expensive and time-consuming to be feasible. The application of IT software innovations to automate and centralize is required to remedy this.

On the other hand, central devices that manage and monitor the cloud will need to be physically secure, extremely durable, and capable of working with ancient serial technologies without a hitch. Regardless if the cloud is built for intelligent transportation systems or for a residential neighborhood smart grid, the physical security of each node in the IoT will require strong protections against vandals, inclement weather, and other environmental challenges. At the same time, the non-trivial problem of interoperability means that 40 year old IA technologies will need to be integrated. Current industrial systems rely on a wide variety of incompatible protocols and interfaces, and establishing Ethernet communications linking, say, multiple, mutually incompatible serial subnets(both ancient and modern) to a remote Ethernet server across a mix of wired and wireless interfaces is a big engineering hurdle by itself, and that is even without taking account of legacy proprietary systems.

Finally, there is the question of how to guarantee solid, dependable AAA and data security during day-to-day operations: without strong, authorization, access, and network accounting controls with secure encryption integrated into the hardware itself, a massively distributed system is only a massively dangerous liability, a disastrous failure waiting to happen.

From a system integrator or industrial supplier’s perspective, these challenges are fundamental problems that must be overcome long before any planning or development of strategic business solutions can take place. Unfortunately, the alignment of these challenges straddles the divide between IT and IA technologies. Setting up and linking automation protocols and mass deployments of edge devices are common enough problems for industrial automation engineers, but for IT engineers they are esoteric disciplines with hidden solutions. In contrast, securing networks with firewalls and AAA servers, linking remote sites over VPNs, overcoming problems like dropped packets and network node failures, and designing wireless networks with strong redundancy and security are all typical IT problems that IA engineers will find time-consuming, at best. It is this IT/IA divide that Moxa platforms are built to bridge.

Eliminating Cloud Complexity with Vertically Integrated Tech

The first step to tackling these problems is getting the right hardware for the job. An industrial-grade cloud solution will need hardware support that’s durable, physically secure, and has broad, flexible connectivity choices that make it easy to deploy and maintain the network. Utility companies, traffic systems, or telcos—each a good example of industrial clouds—will require devices that may be conveniently connected to a network and then quickly and reliably secured, configured, and tested. This frees the engineers who create those clouds from the rote, grunt work of coding what is already known, allowing them to instead focus on the creation and innovation of cloud infrastructure.

At the physical-to-connectivity level, a cloud solution is essentially three distinct layers, with three distinct knowledge sets: a massive array of edge devices deployed in the field, a core control center that receives data for analysis and response, and the wired/wireless connectivity layer that connects the two, running from the edge right on up to the periphery of the central control. Of these three layers, the most difficult to deploy and manage is the connectivity layer: it may include both cellular and 802.11 communications interfaces, as well as wired TCP/IP connections, or even serial connections at the edge. It may be a strictly private network, or it may require communication over the open Internet, using a VPN. Connecting the edge devices may involve the use of incompatible protocols like fieldbus, PROFINET, CAN bus, or RS-485, or they may all be happily integrated using Ethernet and IP.

Such complexity straddles the MAC, networking, and transport layers. That’s why integrated platforms developed for cloud applications must start with prescient hardware designs.

Seamlessly Integrating Every Node on the IoT

After hardware considerations, the next step is building a software environment that facilitates deployment, monitoring, and management. At a network’s edge there will be input/output units, gateways, embedded computers, and HMIs (small or large) that need to be linked to one another, and then back to the control center. Deployments involve not only getting the device connected to the network, but also setting up I/O addressing, alarms, data logging, event-driven triggers, security protocols, and integrating incompatible communication interfaces. The end goal is to keep each individual device as transparent and unobtrusive as possible, and it is here that targeted software automation will significantly cut deployment and maintenance work.

Consider a remote station, where an embedded computer wirelessly connects to a remote control center, while being locally set up as a data logger and event-driven smart control for an array of temperature, wind, and humidity sensors linked to a residential power inverter and an advanced power meter. This is illustrated in the diagram below.

The computer serving this station will need to be compact, have the lowest power requirements possible, hardy enough to withstand harsh environmental extremes, able to be quickly and confidently secured, and—most importantly—equipped with innovative software automation that simplifies administration from the remote control center, while remaining flexible enough to carry higher-level cloud services customized for the solution into which it is integrated.

Moxa’s UC-8481 RISC computers are fanless machines with sturdy shells, highly efficient / low-power RISC processors, and Linux/GNU operating systems. These universal computers offer system integrators and private industrial suppliers rugged, reliable, and easily adaptable edge platforms with strong software optimizations already in place. Yet they offer much more, as well: customized libraries that provide APIs and service routines to aid in the rapid coding of custom communications applications, customized MIB files that expand SNMP monitoring and control, and innovative utilities like Moxa Smart Recovery, a BIOS-level recovery solution that may be fully automated, for both scheduled and emergency recoveries, whether initiated by remote control or locally, by edge devices themselves.

On the other side, core computing solutions will be needed to serve the master system, and this will require rugged, high-performance, easily-managed computers that are capable of reliably fielding a wide variety of edge stations. Moxa’s DA-682A platforms are powerful enough to serve large networks, and come with all of the software features mentioned above as well as Synmap, a software virtualization that adapts SNMP to a high-level process control and monitoring interface. Lastly, in between the core and the edge lies the vast space of the communications layer: switches, routers, and computing stations, wired and wireless, long-range and short.

Obviously, when building cloud solutions, interoperability must remain a key consideration from the very outset. Devices which use open protocols and technologies in a way that eases the work of deployment and connectivity management are much preferable to devices which force engineers to get down into the dirty work of low-level setup and integration. This is where Moxa devices excel, giving IoT engineers a full selection of networking and computing solutions, filling every niche from edge to core.

By combining Moxa computing platforms with our latest software and IA networking innovations, it is now possible to create a virtualized layer on which end users and system integrators may conveniently and quickly build a fully-realized cloud solution. With software innovations like MXview, Active OPC Server, MXconfig, Smart Recovery, and Synmap, Moxa platforms are automated in ways that vastly simplify the deployment, setup, and management of intermediate networking and edge devices. With embedded computing platforms for both core servers and edge stations, as well as IP cameras, switches, routers, gateways, and remote I/O units designed with market-specific compliance for train, marine, and hazardous environments, industrial cloud engineers can now call on Moxa to help them set aside the laborious work of device integration and setup across the entire hardware layer.

With Moxa software automations, you gain a platform that eases every aspect of protocol compatibility, interface interoperability, data acquisition configuration, diagnostics, and addressing, making the connectivity layer as transparent as possible so that you may, for all practical purposes, happily ignore it­­­­­. Whether you are configuring thousands of devices for both processes and network layers or zooming in to view the status, accounting data, and configuration of individual devices at the farthest edge of the cloud, Moxa makes your job easier, intuitive, and more efficient than you ever thought it would be.

To learn more about Moxa cloud computing solutions for IoT, please visit our featured microsite or download our white paper.

Back to index