Author: Aaron White, Regional Director, Middle East at Nutanix
This is the second part of a 2-part series of articles. You can read the first half here.
Jump-Starting IoT Efforts
Many companies have difficulty identifying their best opportunities for IoT. Your goal should be to identify the places where IoT can make the biggest difference:
-Where would having better data help your organization with decision-making or operations planning?
-Do you have the data analysis tools in place to analyze the IoT data once you gather it?
-Which manufacturing processes are the most troublesome? Could IoT data help fix those processes?
-Which front office processes are the most troublesome?
-Could better or more complete data address those issues?
Once you have identified a list of processes that could be improved by IoT, the next step is to identify the ones that pertain to equipment that either already has or can accommodate the necessary sensors and instrumentation. Most importantly, in order to accommodate IoT needs, infrastructure flexibility is essential. If the current IT is built on infrastructure that’s complex, expensive, and rigid, it will not be easy to accommodate new IoT demands. A flexible, agile cloud-based approach will make it easier to adapt to IoT and other new resource demands—both on-premises and in the cloud.
Manufacturing and Artificial Intelligence (AI)
Industry leaders are taking notice of the potential of AI to transform manufacturing. According to a recent report from Infosys, companies want to use AI to further automate manufacturing in order to increase productivity, minimize manual errors, reduce costs and eliminate the need for humans to perform repetitive tasks. These benefits apply across a wide variety of AI use cases from the front office to production facilities.
AI Use Cases in Manufacturing
There are a number of AI use cases that manufacturers are already targeting. AI has the potential to help across your entire company, facilitating fraud prevention, predictive ordering, and opportunity assessment. Below are some:
-Adaptive manufacturing – Today’s customers want products customized to their needs or taste. AI technologies are the key to making today’s rigid manufacturing and assembly line processes more flexible and able to adapt to changing demands quickly and with far less human intervention.
-Human/robot collaboration – For some tasks, humans can’t be replaced. Artificial intelligence and improved sensing capabilities will enable new or updated robots to work more closely with humans, quickly learning new tasks as needs change and making the whole process more adaptive.
-Quality control – Performing anomaly detection on hundreds of units in seconds, rather than hours, enables manufacturers to identify and resolve production failures before expensive delays pile up. Applying AI to quality control not only accelerates production, it can improve quality and reduce reliance on manual human inspections.
-Streamlined supply chains – AI can help make sense of supply-chain data, identifying hidden patterns and improving performance across diverse areas including warehousing, transportation, production, and packaging while also allowing operations to adapt more quickly to market changes.
-Predictive maintenance – Most manufacturers do equipment maintenance on a set schedule that doesn’t account for actual operating conditions. By analyzing equipment log data for anomalies, AI can enable companies to adapt maintenance schedules to actual needs and identify potential problems before failures occur.
Operationalizing AI for Manufacturing
A good rule of thumb for AI projects is: don’t reinvent the wheel. There is no reason to spend your time developing AI models to solve problems that someone else has already solved. On the other hand, you won’t achieve a competitive advantage using the same tools as everyone else, so focus on innovating in the areas where you can differentiate your company. There are two important parts of AI where IT infrastructure decisions are critical:
-Managing data – The first step is to collect data efficiently from the source: IoT sensors, customer equipment, supply chain partners, etc., and stream that data for processing— usually to a datacenter or in the cloud.
-Building and operating AI training clusters – If you are creating and training deep learning algorithms, that requires a training cluster equipped with GPUs that can process data in parallel.
Manufacturing and Advanced Automation
The opportunities for automation across various types of manufacturing are almost endless. Some companies are even evaluating the potential for factories that are 100% automated, with environmental conditions optimized for machines not humans.
IT Infrastructure to Support Automation
The entire operational environment—including IT—is going to become a lot more dynamic than it has been in the past, which has a number of important implications for IT:
-Modernize infrastructure. If you are currently running a lot of legacy infrastructure, you will need to modernize to the greatest extent possible to reduce technical debt, increase agility, and free up resources to focus on new value-added projects.
-Get ready for the edge. As your company rolls out IoT, AI, and automation projects, you will need a lot more IT infrastructure in edge locations such as production facilities, distribution centers, and remote offices. Many of those locations may lack onsite IT staff, making centralized management capabilities and ease of use essential.
-Start by automating IT. To keep up with the increased dynamism and scale of your automation efforts, you will need the ability to automate IT tasks. Consider a private cloud to allow stakeholders to access IT resources and services on demand via self-service.
-Leverage the cloud. Look for opportunities to take advantage of resources in the cloud where it makes sense, including public cloud services, specialized cloud service providers, and SaaS partners.