Last week, Amy Williams, myself, and a group of experts from industry, academia, and nonprofit organizations came together to tackle an important topic facing the biomanufacturing industry: How do we create a fully digital supply chain?
The benefits were quickly agreed upon. At the forefront of most attendees’ minds was the COVID-19 pandemic. Our timely response to the outbreak—with the rapid adaptation, development, and scale-out of a vaccine—was achieved through sheer brute force, supported by government funding and facilitated by close cooperation with regulatory bodies. Many key players involved in the pandemic response have undoubtedly invested in various digital transformation initiatives, and Skellig is proud to have played a critical role in many of these. However, no one can claim to have achieved a fully digital supply chain. Jonathan Wise, CTA[1] at CESMII, set out to uncover why efforts to create a digital supply chain have failed to scale.
Facilitating the Digital Supply Chain
The team at CESMII found that if they followed the flow of data from raw material through supplier, manufacturer, distributor, retailer, and customer, everyone had invented their own data models. Efforts to create an enterprise semantic hierarchy, such as a unified namespace, have facilitated interoperability at the shop floor level but fail to capture the complexity of the entire supply chain. Moreover, it does not inherently enforce any standard data models, which perpetuates the difficulties involved in sharing data across multiple enterprises.
Jonathan and the team at CESMII proposed an alternative with two critical features. The first is what they’ve termed Smart Manufacturing Profiles. These are effectively data models maintained by various standards bodies, including the OPC Foundation. They define the objects and interfaces for smart devices and equipment but can easily be extended to describe any interface where a reliable exchange of data needs to be made, such as digital certificates of analysis. These data models are stored in the Smart Manufacturing marketplace and repositories like GitHub, where they can be referenced.
The second critical feature is a standard API interface that CESMII is developing for requesting data and information from these Smart Manufacturing Profiles. The proposed solution leverages existing cloud hyperscalers such as AWS, Azure, or Google Cloud to persist the data in a graph database. Suppliers make standard REST API calls requesting a specific smart manufacturing profile. If a match is found, the API responds with a short-lived token that allows the supplier to directly access the data in the system.
Leveraging the SM Interoperability Platform for Digital Certificates of Analysis
Leveraging the excellent work already put forth by BioPhorum’s Digital Integration of Sponsor and Contract Organization (DISCO) program, we can extend the Digital Certificates of Analysis use case to highlight the proposed architecture. First, we would have the Smart Manufacturing (SM) Profile, which might define elements around BatchID, SupplierName, SampleID, TestName, TestResult, and so on. In the diagram below, this SM Profile is labeled “B.” This ‘payload’ of data would be assembled by Company A, potentially using a DataOps platform connected to various shop floor systems. The data would then be landed and stored in a location of their choosing, preferably one that is easily accessible to other consumers. Company B, upon receiving the physical material, would submit a request for the Digital CofA to Company A’s Orchestration service on their Smart Manufacturing Interoperability Platform. The API will look within its database of all the potential SM Profiles that Company A has available. If a match is found and the transfer is allowed, the API will return a token that will allow Company B to request the data matching specific values (e.g., BatchID = 123 in the example below). Lastly, all transactions will be stored to a ledger for traceability.
Challenges and Looking Forward
It is worth noting that CESMII is actively working on the development of the API that would facilitate this exchange of data. Still, several major gaps in the industry were identified by the attendees of the workshop. Most notably, the data models are lacking in the biopharmaceutical manufacturing space. This seems like the perfect opportunity for groups such as CESMII to collaborate with other existing industry groups, such as BioPhorum’s DISCO workstream, to not only define the standard data model but also find a standards body holder who can maintain and iterate these data models.
For the industry, the benefits are clear: more robust supply chains that can quickly adapt to surges in demand, such as pandemic situations or other geopolitical issues that threaten fragile aspects of the supply chain where there might be an over-reliance on sources from specific countries of origin. However, many challenges still remain, and input from across the industry is needed. CESMII is actively organizing a follow-up in-person workshop to continue pushing the conversation forward, and we look forward to continuing to contribute to this initiative.