The game changers: the next wave of industrial control trends
‘Horizontal’ IT technology game changers
Based on our ongoing research and industry conversations with companies such as PTC, Carriots, GE, Intel, ARM Holdings, AT&T, IBM, Verizon, Cisco, u-blox, Kontron, Rockwell Automation and Schneider Electric, we recognize the convergence of several key hardware, software and delivery model technology trends that will affect the pace and scale of a new round of investment in industrial control and automation enablers. First, there are many key IT-centric enabling technologies that will be critical to the development of industrial process control evolution through industry scale, cost advantages and security attributes. These enablers have the advantage of wide application across consumer and enterprise markets and massive industry scale, driving costs down precipitously.
Embedded computing/advanced sensor technologies. Advances in embedded computing systems will remain a core game-changer as M2M and SCADA systems evolve to IIoT variants. Embedded computing systems have long brought dedicated computing functionality into connected industrial systems – generally designed to support a specific function or set of functions. Embedded computing systems are powered either by ordinary microprocessors using separate circuits for memory or peripherals or use integrated microcontrollers, which provide all-in-one functionality, thereby reducing costs, size and power consumption. System-on-a-chip (SOC) approaches can host all electronic components on a single chip that can operate for years on a single battery. In addition, the increasing availability of better, smaller, cheaper and smarter sensors will allow machines to feel, see, hear and react to their operating environment. These sensors pair up with tiny microprocessors to form microelectromechanical sensors (MEMs) that have the ability to work and make decisions autonomously or to work as part of a larger connected system.
Key vendors: Intel/Windriver, Kontron, ARM, Integrys, Synopsis, Eurotech, InvenSense AMD, Freescale, STMicroelectronics, Texas Instruments, Kionix, Qualcomm, MicroGen Systems, Qualtre Inc, Semtech, Advantech, NXP Semiconductors, Imagination Technologies.
Ethernet/IP/Wi-Fi/gateways. The proliferation of industrial Ethernet/IP networking and IoT optimized gateway infrastructure provides the economies of scale of a general IT technology with the data format requirements of industrial equipment. Advanced gateway designs provided rugged aggregation points for sensor data, which can also provide proxy security services for downstream things with limited on-device countermeasures. The IPv6 standard brings exponentially larger address spaces and improved security capability. The availability of unlimited addresses changes the conversation about addressing schemas for industrial networking architecture. IPV6 can easily support every human and machine on the planet. Industrial 802.11 b/g/a/h/i access points are becoming ubiquitous in the industrial setting.
Key vendors: Cisco, Siemens, Digi International, CradlePoint, Intel, ARM, Eurotech, B&B Electronics, Moxa, Advantech, Lantronix, Multi-Tech Systems, Sierra Wireless.
Cellular broadband and unlicensed wireless. In the past five years, the pricing for 2/3/4G cellular connectivity has fallen drastically due to hypercompetitive market conditions and strong motivation on the part of MNOs and MVNOs to win M2M/IoT business around the globe. In just the past 12 months, LTE module pricing has fallen by more than 50%; 2G modules can be had for less than $10 when purchased at volume. At the same time, chipmakers are offering multiple connectivity options directly on chipset hardware. These forces have made cellular connectivity a viable alternative to legacy architectures such as frame relay, ISDN, wireless mesh and power-line networking (PLN). While MNOs will be eager to re-farm 2G bandwidth for 4G, the market has produced ultra-narrow band (UNB) alternatives from outfits like SIGFOX or driven by the LoRa and Weightless alliances that plan to offer purpose-built networks for machine connections that don’t require broadband bandwidth and support super-low cost and edge power requirements. At the local area network level, Wi-Fi, ZigBee, Z-Wave and Bluetooth LE networks provide ample connectivity options over unlicensed spectrum over very low-cost radios.
Key vendors:AT&T, Vodafone, Sprint, T-Mobile, u-blox, Telefónica, Deutsche Telekom, Orange, Telit, Gemalto, Sierra Wireless, SIGFOX, Actility, Multi-Tech Systems, Intel, Freescale, ARM, Qualcomm, KORE Wireless Group, Aeris Communications, Wyless, Numerex, CalAmp, Kyocera, Novatel Wireless, Sagemcom, SIMCom Wireless Solutions, Huawei, NWave Technologies, Semtech, NXP.
Cloud computing/storage/network models. The availability of high-quality, virtualized, on-demand compute, storage and network resources in public, private or hybrid configuration has forever altered the enterprise IT infrastructure model for deploying and supporting enterprise application workloads. The game-changing economic and scale advantages of the cloud model align perfectly with the requirements of IIoT and are arguably the most critical enabler for its viability. The important caveat to keep in mind is that data storage and ownership are sensitive – the degree of comfort with housing mission-critical data outside the enterprise on public cloud services will vary (industrial manufacturing companies tend to be very conservative in this regard).
Key vendors: Amazon, Microsoft, CenturyLink, IBM, SAP, salesforce.com, Oracle, Verizon, Google, Rackspace, HP, NaviSite.
IoT connectivity and data platforms. With increasing demand for reliable, ongoing and secure connectivity to connected assets, a new crop of software platforms and PaaS alternatives have come online from specialists and IT heavyweights designed to reduce the cost, time and complexity of managing connectivity, machine data storage and processing, application development, security and integration with existing IT systems. These products provide compelling alternatives to the internal IT-driven approach favored by some of the early adopters of M2M.
Key vendors: GE, PTC, SAP, salesforce.com, IBM, Oracle, Carriots, Exosite, EVRYTHNG, LogMeIn/Xively, Jasper, Stream Technologies, Ericsson, Aeris, KORE, Vodafone, Verizon, RTI, Octoblu, BlackBerry/QNX Software Systems, mbed (ARM), Software AG, Amazon (2Lementry), Bosch, Wyless, Ayla Networks, Arrayent, Microsoft.
‘Big data’ software tools. Without the tools to make sense of the impending blizzard of machine data, the IIoT hype will prove to be all bark and no bite. The great innovation possible in IIoT involves bringing massive internal and external data sets together to uncover new insights. Luckily, the demand for data tools has driven massive research investment that has produced innovations such as Hadoop for massively parallel processing of data sets. In-memory database systems are also promising for helping business leaders make decisions based on IIoT in real time. Data scientists who can combine deep vertical sector expertise with data sciences will be able to set their own market price because these skills will be more valuable than many others in organizations. The lack of these skills in the manufacturing sector presents a significant challenge for OEMs and a service opportunity for vendors and integrators with these skills.
Key vendors: Splunk, Software AG, SAS, Oracle, MongoDB, TIBCO, IBM, SAP, Cloudera, Microsoft, Informatica, Teradata, HP, Cisco, DataTorrent, Glassbeam, ParStream, GroveStreams, Snowflake Computing, Infobright, CRATE Technology, SQLstream, FeedZai, DataTorrent TempoIQ, Lokad, jKool, Loggly, Logentries, X15 Software, Cloudera, MapR, Hortonworks, Pivotal, MongoDB, DataStax, Treasure Data, Accenture.
Intelligent edge networking. Innovation here represents the intersection of big data and next-generation network infrastructure. New network designs eschew the logical data path that follows data packets from the edge all the way back to the datacenter. A new crop of technologies will combine real-time analytics software directly into the router path at the network edge. Cisco and others are calling this ‘fog’ computing. Edge intelligence brings significant value in a scaled IoT implementation on two axes. First, policy logic dictates that only data that ‘matters’ is sent upstream, reducing the amount of data traffic that must flow from edge to core. Second, policy decisions can be made at the edge – for instance, immediately shutting down equipment if a certain alarm condition occurs. The data volumes and response times anticipated in large-scale IIoT applications will call for edge devices at different layers to have greater autonomy and more intelligence; for instance, by using business logic to only send exception or alarm data back to the datacenter based on established business logic, thereby saving large volumes of data traffic that is captured – but is not particularly useful or doesn’t require an immediate response.
Key vendors: Cisco, Ericsson, Juniper, Nokia, Intel, Alcatel-Lucent, HP, B+B SmartWorx.
Industrial automation-specific game changers – near term
Advanced industrial automation technology. The latest generation of industrial control technologies is getting smarter and more efficient. Take, for instance, the ubiquitously deployed variable frequency drive (VFD), a control element that is used in electromechanical drive systems to control AC motor speed and torque by varying motor input frequency and voltage on mechanical systems such as pumps. Until this year, VFDs have not been connected products – if a drive had a fault, it required someone to manually intervene and physically read an error code from the machine itself. Schneider Electric recently released a new connected VFD that comes equipped with Ethernet/IP connectivity and its own embedded Web server. Now the product is far more intelligent and can work in conjunction with power-monitoring applications that bring game-changing granularity to the drive’s performance and power consumption. Given that nearly 25% of the globe’s electricity consumption can be attributed to industrial equipment, the potential savings for plant owners and the environment are particularly compelling.
Key vendors: ABB, Omron, Schneider Electric, Rockwell Automation, Emerson Electric, Siemens, Opto 22, Honeywell International, Johnson Controls, Texas Instruments, Datalogic, Yokogawa Electric.
Mobile applications for industrial control. An often-overlooked trend within the industrial control market is that industrial automation equipment can now be monitored and controlled via smartphones and tablets running apps used by plant engineers and line staff. The latest generation of industrial control equipment will be mobilized and come with apps that provide many of the productivity enhancements that enterprise mobility has already brought to non-production workforces. These apps can bring tremendous value to industrial operations, enabling decisions on the fly with real-time information. One of the major challenges to overcome is a lack of common accounts with Apple or Google such that employees can seamlessly download these applications onto their personal devices. Devoid of these accounts and clear policies, employees are left to wonder if they can be reimbursed for purchasing these applications for use on their personal device.
Key vendors: Apple, Google, Microsoft, Rockwell Automation, Schneider Electric, Siemens, Opto 22.
Industrial automation-specific game changers – longer term
Predictive analytics/machine-learning software. While it is generally agreed that machine learning is still a nascent concept, the application of predictive analytics to industrial machinery is already making some progress. By connecting their industrial machinery with big data and predictive analytics, companies such as GE can predict failures before they actually happen, reducing downtime and improving the level of service delivered. This is truly game changing to all parties involved. Plant managers will be able to understand the dynamics of their ‘systems of systems’ to make better decisions on plant design and operations. Machine learning aims to imitate human learning through mimicking learning behavior and decision-making functions at far greater scale and speed. Artificial intelligence and machine learning will be foundational components of autonomous factories that will discover new knowledge and make decisions without human intervention – an exciting, and scary, thought.
Key vendors: SAS, SAP, Predixion Software, IBM, Rockwell Automation, Microsoft, Oracle, RapidMiner.
Wearable technology. The past year has seen moderate adoption of wearable computing technology in the consumer sector, although the recent decision by Google to hit pause on Google Glass can be viewed a setback for that category. Wearable technology is predominately made up of wrist-worn devices such as smart watches, fitness trackers and head-up display (HUD) devices such as the now discontinued Google Glass. There are a number of companies focused on bringing both technology categories into the industrial setting. It won’t be long until we see plant workers accessing information via HUD devices, relaying live video feeds from mechanical situations back to decision-making authorities.
Key vendors: Google, Apple, Samsung, Qualcomm, Kopin, Vuzix, Motorola, Laster Technologies, CrowdOptic.
Business drivers and inhibitors
There are a number of drivers – technical and non-technical – that we expect will drive the investment in IoT-centric industrial automation at a significant pace. There are also significant hurdles spanning technology, process and organizational cultures that must be overcome.
Non-technical market drivers include the following:
Government programs. In Germany, the government is active in its promotion of what it calls Industry 4.0 to help industries harness the intelligence generated by the IoT to optimize processes, increase efficiencies and spur innovation. The European Union has committed over €1bn for ‘factories of the future’ research. In the US, the Smart Manufacturing Leadership Coalition has focused on promoting ’21st century smart manufacturing,’ involving best practices including reference architectures to support the integration of OT and IT environments.
Reduced cost and increased efficiency. While we have made a case that pure cost reductions are sometimes not enough to tip the scale in favor in replacing legacy OT technology, the ability to save money is still a major benefit of technology upgrades. Smart, connected products can help plant managers make products with better quality, in less time and using less energy. Once the product leaves the factory floor, ongoing connectivity to physical products provides game-changing insights and value for customers and suppliers alike.
Increasing competitive pressure to get connected. As the IIoT becomes real, companies that eschew the technology will find it increasingly difficult to compete on a global scale and will find that they are vulnerable to displacement by upstart competitors. The quality and time-to-market gains that can be realized by a connected plant using the latest innovations in industrial automation and IT compared with existing setups translates into compelling competitive advantages.
Availability of new business/servicing models. Industrial automation vendors with hyper-connected offerings may choose to sell them in new ways – for example, by moving from traditional capital financing models to operational models that turn product sales into ‘as-a-service’ models – so-called ‘servitization’ strategies. These new business models can bring technology to small- to-medium-sized OEMs that may not have access to the capital required to invest in the latest and greatest technologies. For suppliers of industrial machinery, connected products provide ongoing feedback loops into product design and usage patterns; in this way, the commercial market becomes an ongoing test bed and a driver of future innovation. Smart, connected products can be packaged in a variety of business and support models.
The IIoT carries significant constraints to adoption. Some of them are legal, some cultural and some technical. These inhibitors are partly responsible for the poor level of adoption of IoT technology in the manufacturing sector to date.
ROI challenged by massive investments tied to legacy infrastructure. As might be expected, legacy infrastructure investments are significant inhibitors of new technology without a compelling new value proposition beyond cost savings. Network infrastructure based on frame relay and ISDN can be found in legacy industrial automation networks, while serial networking protocols such as Modbus are widespread among legacy I/O devices that still do their job effectively. A key value proposition that suppliers bring to the table is the ability to overlay new infrastructure into the legacy environment to bring this older technology into newer industrial automation systems, thereby protecting the original investment and layering in additional value by allowing them to participate in a modernized IT environment. Retrofitting existing technology with sensors is an easy interim step until the product lifecycle calls for a new industrial machine that comes embedded with IoT technologies.
Challenge level: High
Cultural issues. The cultural issues involved with complex technical evolutions are directly related to the legacy infrastructure challenge. The convergence of OT and IT brings as much cultural challenge, if not more, than the actual job of replacing old OT gear with new IT. These two camps have typically worked together at arm’s length – if at all. While OT engineers bring deep expertise in process control technology, power management and high availability, they lack training and practical experience in basic IT security infrastructure, patching and networking. IT can build and scale a secure network, but lacks the contextual understanding of what is required in the operational environment, which is different and often mission-critical. Several interviewees we spoke with have said that ‘OT and IT guys don’t like each other.’ The only way to overcome this challenge is to organize OT and IT under common leadership, with a common purpose and appropriate cross-training and common language and governance. This is perhaps the largest challenge holding the industry back at the moment.
Challenge level: Very high
Data ownership and rights. The benefits and resultant growth of IIoT will not be realized without making good use of the data exhaust blowing off connected machines. As more systems come online, the network effects of community participation come into effect and will create step-changes in new insights and value. But how is this data to be shared, and with whom? Who actually owns it all? Product manufacturers will surely want to own some of the data flowing from their machines, but the customer will likely have some rights to it as well. These are thorny issues that are not easy to answer and will require regulatory intervention in a number of scenarios. Currently, connected machine manufacturers are addressing these issues on a customer-by-customer basis – some customers care ‘a lot’ about retaining ownership of their data, others will give up some ownership to improve their processes.
Challenge level: High
Security. The absolute best excuse to remain on the IoT sidelines is the very real threat of security risk and exposure. The attack surface expands exponentially for IT security leadership with the explosion of connected machines. Every IP-addressable physical asset will become a potential attack target or point of ingress for hackers. Organizations in industrial control, manufacturing, IoT and other similar verticals have been worried about attacks against their purpose-built systems for some time. The highly publicized Target data breach was widely rumored to have emanated from the company’s connected HVAC system. CEOs and CIOs are rightfully concerned with opening their firms’ reputations and brands to known and unknown threats – with IIoT, there are plenty of both to worry about.
Challenge level: High
Lack of skills. There are currently over three million unfilled manufacturing jobs in the US. As a nation, the US needs to re-educate future computer science and engineering graduates that there are plenty of hard IT challenges to be solved in the fields of manufacturing operations. The repatriation of the manufacturing sector to the US has been fueled by a number of factors, including cheap energy and better tax advantages. This trend will only be slowed if highly skilled data scientists and IT specialists cannot be attracted by exciting business challenges and the opportunity to apply the latest and greatest technologies to revitalize the domestic manufacturing sector. This shortage of skills opens opportunities for suppliers to lend their own expertise either directly or ‘as a service.’
Challenge level: High