Become a member of our mailing list to get the latest news, articles, and insightful content.
Subscribe to our Newsletter
Thank You for subscribing with us. We sent you an email regarding this.Subscribe Now!
Industrial IoT is clearly one of the top areas of interest among CIOs and business line owners because it offers two main business advantages; The first is hard operational cost savings and efficiency through dramatic processes improvements, and the second is enabling new business models with new revenue streams and competitive advantage. Consider, AVIS CarRental company, leveraging an IoT strategywith Telematics and sensors, for a seamless car rental experience. AVIS “expects to save $100 million by 2020”, and gain competitive advantage from a new IoT managed system, driven by business use cases and outcomes.Stanley Black and Decker, in its manufacturing plant in Mexico, uses location based tags and analytics to improve machine production line efficiency from 75 percent to 95 percent, and improve product quality while reducing cost of manufacturing and inventory.
Industrial IoT Adoption Challenges
As enterprises are starting to see the tremendous benefits, some of the key challenges (both in business and technology) that CIO’s face as they embrace IoT are the follows:
- What problem to target: Which use-cases offer the most value for the Enterprise?
- ROI: For those use-cases how to calculate the ROI and what ROI models to use?
- Technology Leverage: How can existing and legacy technology infrastructure be leveraged for IoT driven systems? Many Industrial systems are managed by Legacy operational systems, with huge existing investments, and a "rip-and-replace" makes no business sense.
- Technology Platform: There are hundreds of IoT Platforms, so how do you go about picking a platform for your use-case?
" Edge computing is really becoming a primary early-stage design consideration based on a lot of Industrial and Enterprise IoT implementation"
Business of Industrial IoT
Business outcomes need to drive the IoT adoption. There are two methodologies that I have seen used to build a business case for an IoT initiative. The first, which a lot of early adopters use, is for a business function to run with a pilot with a specific,well-defined business outcome, and a managed scope of work. One such example is getting data from an existing SCADA system, and using it for measuring something new, like the energy consumption or machine friction trends. The second, more strategic path, is based on a facilitated workshop-led program, where IT and business stakeholders work through a list of the most important strategic, financial, and growth goals for the enterprise. Once the use cases are prioritized, a business model and ROI model are then built to align with the business goals. The ROI models for IoT are based on a combination of cost savings and new revenue streams, either from associated or new business models. For example, one enterprise has built their IoT business model for an automated predictive machine maintenance system, based on a ROI of higher machine uptime, and projected energy savings from effective operations.
Technology Adoption of Industrial IoT
Technology Adoption of Industrial IoT
1) Enterprises have invested into existing systems, and many industrial systems heavily rely on legacy control systems for its operations and control. So how CIO’s should think about adopting Industrial IoT solutions?
2) The I-IOT adoption strategies being incorporated by many enterprises varies in size and scope, but mainly fall into three solution groupings:
3) Enterprises are leveraging existing technology, especially in the area of Data Analytics from machine control systems melded to cloud-based infrastructure.
Early adoption of new IoT platforms (GEPREDIX, MSAzure, AWS IoT) that compliment (and will eventually replace and merge) many “home-grown” standalone IT systems.
A combinational approach of integrating existing systems with new IoT Platform technologies.
Most Industrial IoT projects leverage existing systems, including Legacy Control systems, Data Analytics, Cloud infrastructure and Mobility services. What gets added is an overlay of additional sensors for data-sensing, processing on the Edge, and processing in the Cloud. On the Edge, computing and networking has become cost-competitive so that processing of data and even operational Analytics can now happen on the Edge, while only Aggregated data is sent over to the cloud for learning and summary analysis.
While these adoption patterns make business sense, they also create some key technology concerns for enterprises as it involves a lot of retrofitting and integration. These are around Security, Operations and Scale.
Security, Operation and Scale - How doesan enterprise manage the assets(machines and sensors) which can degrade,fail,or deprecate? How does the enterprise manage patching vulnerabilities to a huge estate of sensors? Additionally, how can they ensure machines (that can now be controlled through exposed API’s) are not tampered with in error or malice, and the data that is generatedexploitedmaliciously?The third area of concern is around scalability. A solution might start with small amount of data and processing, but quickly mushrooms into large data and processing needs, and integration with multiple systems. How do you setup a system that can scale over time, without significant re-architecture, and expensivereplacements? How do you ensure the platform is future-proof, and can handle future protocol changes?
New Ways of Thinking: Top IoT Adoption Patterns
These concerns need to be outlined early in the architecture and design lifecycle, with a clear plan for how to address them in initial pilots, and expand them over timeto avoid costly rip-and-replace scenarios as projects grow, andout-of-control costs as the consumption (including the network, cloud, analytics) exceeds original intent and quickly erodes business benefits. So how do you address these concerns? Here are some recommendations:
• Deployment strategy: Several scenarios (such as low latency Analytics and Machine learning) require data to be processed close to the data source. Some critical scenarios include use-cases where large amounts of data can be generated by machines and should be handled locally, scenarios where the communication infrastructure to the cloud is not optimal, and scenarios where there is sensitivity in where the data can be processed. In these scenarios, Edge computing is the way to go. There are several Edge computing solutions rapidly emerging in the market, so computing, analytics and machine learning can be performed on the premise, and any local decisions implemented. Data that is needed to be aggregated can then be sent to the cloud or private data center for further analytics.
• Security and Governance: Like the old adage, “with great power comes great responsibility,” security and manageability is one such multi-faceted problem. It needs to be addressed with an end-to-end approach. Firstly the physical assets need to be protected (includingthe machines and related assets, sensors on those assets, networking gear, mobile devices, and servers hosting the services, are all part of the infrastructure). Secondly,the data that is generated from those assets needs to be protected, including data in-transit, intermediary devices like brokers, edge servers and cloud servers, as well as data at rest.Finally, applications and end-points that access that data need to be secured. Consider a scenario, where a mobile app has the ability to change the operation of a manufacturing machine line. A specialist can troubleshoot the machine remotely, and trigger some remote actions, but it also exposes the feature for malicious activity. Further, manageability is a big deal. The multiple types of physical assets and infrastructure need to be managed for both operational monitoring and maintenance. The proliferation of device types, and multiple protocols to manage those devices, makes monitoring and maintenance a big headache. How I see this getting addressed is a two-step process;the first step is visibility. All the Assets in the ecosystem need to be visible by an operational system. Where assets have the mechanism(such as API’s) to tap into its status, location, etc., they need to be accessed and made visible. The second step is control. There are tools that enable IT to view and set policies on the assets, and track those for optimal performance. As an example, if a type of sensorhas a certain operating range, you need to set actions, such as notifications, based on the optimal performance.
• Scalability: This complements the deployment architecture for an IoT deployment. You need to carefully review your scalability needs, and design the system with the flexibility of moving your processing across the Edge, private data centers, and the cloud. This, in my opinion, is the best approach to build a solution, with the design flexibility to move your processing and storage from the edge to a private data center and the cloud. As an example, consider a scenario where an enterprise is monitoring data from its machine line, and looking for patterns on how the machine functions impact the production of the product from the machines. Every percent point improvement in production volume directly leads to higher revenues and better margins. Architecting the solution so it can be scaled from a small setup to a large deployment is an important design consideration. So, Edge computing is really becoming a primary early-stage design consideration based on a lot of Industrial and Enterprise IoT implementations.
IoT adoption is growing exponentially, with adoption being primarily driven by use cases with hard savings in operational efficiency and dramatic process improvement, and also by delivering new business models- resulting in tremendous competitive advantage. Many enterprises are leveraging existing technology, especially in the area of data analytics, machine control systems, and cloud infrastructure. Enterprises are focusing on business outcomes, starting small and cautiously, and expanding programs based on business results. Leveraging existing infrastructure and systems introduces some risks that need to be addressed early in the lifecycle. IoT deployment models can be very different than traditional IT systems with data, connectivity, security, and scalability driving a hybrid combination of edge and cloud deployment. Manageability and security are a multi-faceted challenge for IoT systems and need to be addressed upfront. Edge computing is emerging as a key enabler for this model to be successful. In addition, the right tools need to be adopted in order to handle security from an end-to-end visibility and manageability standpoint. What IoT initiatives is your business demanding, and have you worked out the foundation capabilities in order to lead you successfully to the new, connected enterprise?