Industrial Data Fabric solution best practices – Part 2, build, launch, and scale

This blog is the second in a two-part series that distills some of the most important best practices to help organizations accelerate the implementation of an Industrial Data Fabric (IDF) solution on Amazon Web Services (AWS) and increase the value it brings enterprise-wide. The best practices are organized by stage, as shown in figure 1 below.

Figure 1. IDF implementation stages

In part 1, we reviewed some of the challenges of IDF implementation and provided high level recommendations for the alignment and organization readiness stage in detail. In this blog, we present insights about the foundational build, launch, and scale & optimize stages. In these stages, organizations follow the architecture vision to build an IDF that yields continually increasing return on investment.

Minimum Viable Product, MVP foundation

1: Plan ahead to implement within a tight schedule and budget

After design and prior to any build activities, your next step should be resolving prerequisites for your target environment based on your proposed design. The prerequisites for IDF implementation may include hardware, network connectivity equipment, data access, and tools, as well as teams who are accountable for readiness (see figure 2 below). Lack of preparation in these areas is the most common reason for implementation blockers and delays while delivering within narrow schedules and budgets since lead time for network connectivity, hardware procurement, and installation may take weeks or even months. It can also often require change board approval to verify that standardization and changes do not interfere with existing infrastructure or production.

Figure 2. Common IDF prerequisites

In addition to identifying these prerequisites, it’s important to know your organization’s change process and involve key stakeholders from enterprise architecture, networking, and plant management in the design phase and obtain commitment for timely approval and readiness. Preparation is key for a successful implementation, but don’t let too much preparation and planning delay the build and outcomes. Start with a relatively simple foundational setup so that delivery teams can quickly build capabilities and iterate toward a well-architected industrial data fabric.

2: Start with minimum viable capabilities to demonstrate immediate business value

One pattern that has proven successful for multiple IDF implementations is to start with a representative and comprehensive use case, with minimum valuable capabilities in scope, implemented at one factory that meets prerequisites for technology standards. This pattern not only establishes the most important architecture elements to maximize reusability but also minimizes risk in the most challenging phase of implementation—laying the foundation. Even more importantly, it demonstrates clear and timely business value to establish increasing stakeholder buy-in as well as momentum to accelerate into future phases, when you will scale out to other factories with similar needs and implement new use cases on top of the foundation.

3: Establish one team and one goal to build faster with greater accountability

The increase of industrial Internet of Things (IIoT), machine learning (ML), data analytics, Industry 4.0, and new business opportunities presented by digital transformation call for a convergence of operational technology (OT), information technology (IT), and cloud technology to digitize business processes. Many times, IT leads plant optimization, but they are not successful without being tied to the plant operations team. The operations team needs IT system data to optimize processes, but they may not have the technology or skills to do so.

A pattern that has proven successful in accelerating IDF implementation requires a less siloed relationship between IT and OT. In fact, building IDF faster with greater accountability requires a one-team mindset and environment. As shown in figure 3 below, the one-team approach combines talent from across OT and IT and works closely with the plant team and other functions (such as infrastructure, cybersecurity, networking, and so on). An executive sponsor from the C-suite should ensure that the one-team’s attention is focused on the success of the initiative alone. Also, the executive sponsor must empower a single-threaded leader (STL) whose focus is also IDF alone.

Figure 3. The one-team approach

The STL leads the centralized team and is responsible for all aspects of the success of the IDF initiative. The STL verifies that all roles talk to and understand each other, gather common value drivers, gather detailed requirements and prioritize features, synchronize with the various organizations to take their road maps into consideration, and escalate risks early.

4: Identify and address skills gaps early

Digital transformation and the IDF technical landscape require new skill sets in IIoT, the cloud, data analytics, and ML to build pipelines for data collection, data quality, and business insights and controls. Skills gaps in these areas will ad/d confusion; slow down the pace; and limit the ability to own the solution, scale, and innovate. However, addressing skills gaps takes significant time and effort. Consider the following three steps when beginning an IDF implementation:

• Identify skill gaps early: Essential skills that are often missing in IDF implementations include a lead architect, who can provide technical guidance on the current and future state and drive alignment across technical teams; IoT engineers, who can build edge capabilities and configure asset hierarchy; and data engineers, who can build data ingestion, standardization, and curation pipelines. These roles and skills are essential to build and manage the solution and must be onboarded from day one. ML roles can be onboarded when data collection begins.
• Create a training plan and secure technology partners to close the skills gap: Address skills gaps by adding resources from your technology partner and creating a targeted training program to upskill existing employees. Take advantage of the training and certifications programs offered by AWS and your partners.
• Continue to assess and upskill: Industrial modernization is a journey where workforce training and development involves continual improvement. When you begin your IDF journey, initial or partial understanding of the technology and competencies required is normal. But what’s important at this stage is having a solid plan to gain a complete understanding of the system and technology prior to production rollout and to continue to assess and address skills gaps to increase the pace of innovation.

Launch

5: Start handover to the sustaining team early

The alignment and foundational build stages can also be led and codeveloped by AWS Professional Services—a global team of experts that can help you realize your desired business outcomes—along with your technology partners and team. Often, AWS Professional Services and your partners assume a “builder” role to develop a detailed design and capabilities for the foundational solution. During the launch phase, the AWS team transitions from a “builder” role to a “coaching” role to help you and your partners to understand, deploy, and own production of the IDF and use cases. The foundational stage is only successful when your IT, OT, and partner teams have a thorough understanding of the end-to-end system to adopt and own production viability of the new solution. To facilitate this, develop a transition plan during the high-level design phase and begin to codevelop the low-level design and build components with the team who will own it.

Figure 4. Teams involved with the implementation stages

6: Conduct production readiness planning

When you reach the 75% percent mark of the foundational build and begin to realize initial benefits of the solution, start the transition to production readiness planning. In other words, launching a new solution in production requires careful planning and 1–2 months of coordinated effort to prepare the infrastructure and data migration; train the local plant, run, and maintenance teams; and establish operational monitoring, change management, and incident and service management. The following are key elements:

• Training and plant readiness: One of the risks observed with IDF implementation is plant readiness with reliable connectivity, end user training, and willingness to move to new tools and processes. Be sure to engage the plant team early to socialize minimum viable product (MVP) capabilities, collect feedback, and take corrective actions. Making process, technology, and tool changes for the plant is not easy, and lack of preparation could delay outcomes. Develop change management strategies to introduce new technology, provide user training, and onboard incident and service management for technology and infrastructure issues.
• Hyper-care and cross-functional commitment: Verify that the central one team and cross-functional teams have capacity planned to provide hyper-care support for production launch, incremental onboarding of new use cases, and scaling to more factories. During hyper-care support, the central one team and cross-functional teams need to troubleshoot, operate the new solution, resolve early go-live issues, assess possible changes, and update run books until the new solution is stable and long-term operational support assumes full responsibility.
• Preparation of technology partners for long-term support: At most organizations, technology partners are critical to help build and maintain the solution landscape. Engage partners during the pilot or foundational build phase, and verify that partners share a similar vision for transformation and provide the required support for the long-term strategy.
• Another AWS Well Architected review: Review gaps and readiness, continuing on from the previous review using AWS Well Architected, which helps cloud architects build secure, high-performing, resilient, and efficient infrastructure. Verify that the architecture, security, operations, reliability, and performance meet production-grade requirements.

Scale and optimize

7: Adopt ongoing practices to succeed
While the initial launch proves solution viability and demonstrates customer outcomes, the real impact and business value come only when the solution is adopted across the entire organization and its locations. The adoption and scaling phase is the hardest part of the IDF journey, but organizations increase their chance of success with the following actions:

• Facilitate continual adoption and innovation: Scale out the foundation to other plant or shop floors with similar needs and implement new use cases on top of the foundation. Centralized OT/IT is instrumental to accelerating adoption of the IDF and to collaborating with business units (BUs) and cross-functional teams. The centralized OT/IT focus is on further foundational build-out when incorporating new plants and on iteration to facilitate new use cases with new data sources and cataloging. Identify opportunities where the foundation can provide immediate value through the data catalog and standardized, centralized access to data, and invite different parts of the organization to consume data for their needs.
• Establish governance and promotion: Governance is an essential part of the transformation process, verifying alignment, guiding innovation, and nurturing the convergence of IT and OT and other BU partnerships. Internal marketing and socializing through lunch and learn sessions and workshops can help increase company-wide awareness and adoption.
• Measure success: Measure success with key performance indicators (KPIs) for adoption and results. The KPIs for adoption could be plant satisfaction, number of shop floors connected, number of pieces of equipment connected, or number of data sources and users trained. Results KPIs include production downtime, employee safety, and others.
• Standardize: Industrial processes require a wide variety of equipment from different original equipment manufacturers (OEMs). Having standardized networking and data collection facilitates smooth scaling as more assets and data sources are added to an IDF solution. Verify that the centralized OT/IT continues to align with enterprise architecture and AWS Well Architected standards to improve security, optimize for cost and performance, drive standardization, extend architectures, and create repeatable practices.
• Make it extensible: Industrial landscapes often have legacy systems, including closed control systems, manufacturing execution systems (MESs), enterprise resource planning (ERP), and others to collect data and monitor and gain insights across production processes. It’s critical that IDF is extensible to legacy systems through edge capabilities and APIs to provide near real-time insights for maintenance engineers and visualizations for managers.
• Maintain an agile culture: Constantly assess your delivery methodology and working agreements, and encourage incremental, iterative, and adaptive approaches to accelerate adoption and deliver outcomes more quickly.
• Uplift talent: Continue to assess key roles and skills necessary for fast-changing business conditions and modern technology environments. Build talent through internal mechanisms and partners.

We covered some of the best patterns and practices in implementing IDF from organizational, people, process, and technology perspectives. These recommendations stem from ongoing collective learnings of successful IDF implementations for manufacturing and industrial customers. Follow the tried-and-tested steps and tailor them to your organization’s needs to succeed in your digital transformation initiatives and accelerate value realization. To learn more, see AWS for Industrial and Industrial Data Fabric (IDF). You can also reach out to your sales or ProServe representatives to explore how to achieve great success with IDF!