BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Blockchain promises several benefits for supply chain applications, among them increased transparency, visibility, governance and auditability.
However, IT managers might want to consider several problems before launching a major blockchain initiative. Blockchain, also called distributed ledger technology (DLT), includes a variety of technology stacks that can address some of these problems better than others. An understanding of these blockchain problems can lead to better discussions with supply chain partners and vendors before launching a blockchain project.
Forrester estimates that more than 90% of the enterprise DLT projects started between 2015 and late 2017 have already been discontinued or will never result in an operational system.
Martha Bennett, a principal analyst at Forrester, attributes some of these failures to starting a blockchain project without due consideration of a viable use case or technical feasibility. It's much better to start by understanding what an enterprise is trying to achieve by using blockchain to improve a supply chain before choosing a particular technology.
Not an incremental update
Adopting blockchain for a supply chain application requires a lot of groundwork. IT managers and developers may be tempted to experiment with a proof of concept to gain a better understanding of the technology, but it won't always disclose many of the underlying challenges in a real working solution.
"Many of the early experiments were designed to be quick and dirty buildouts," said Marta Piekarska, director of ecosystem at Hyperledger, an open-source DLT development project. "However, this approach often doesn't plan for scale, and so, when entering the hard reality of the markets, enterprises see how many factors were omitted or forgotten."
A good practice to avoid blockchain problems is to collaborate with other companies, even those outside of the industry, to identify technologies, practices and business processes that can be adapted to working supply chain solutions. For example, Hyperledger currently works with more than 250 companies with collaborations like the TKI Dinalog Dutch Institute for Advanced Logistics project and the Belt and Road Blockchain Consortium.
Higher computational overhead
Blockchain applications used in supply chains are far more efficient than currencies like bitcoin, but they still require more computing resources -- or overhead -- than traditional databases. This means some of the operations of these applications, such as reading from the blockchain, may be significantly slower compared to those operations in traditional databases.
"Creating a large network to handle an industry will require massive scale due to the complexity of handling potentially hundreds of thousands of distributed ledgers," said Bill Clabby, a research analyst at Clabby Analytics. "I don't think the market can handle a massive scale servicing tens of thousands of vendors and devices at the present time."
Clabby recommends that IT managers consider an approach like IBM's Blockchain, which is built on mainframe technology and scales better than other complex instruction set computer or computing and RISC architectures. "Mainframes are extremely well designed for transaction processing, and that's what blockchain is," he said.
Although blockchain technology could support decentralized applications managed by separate vendors, they most likely won't be as easy to manage as more centralized, industry focused exchanges.
Blockchain application exchanges may not be popular with enterprises at first, much like when music distribution services, including Apple Music, received pushback in the beginning. But, over time, enterprises may have to adapt to these new exchanges, just like the music industry did.
Control essentially plays a bigger role in blockchain versus music because it relates to mission-critical transactions that are vital to the business.
"If the healthcare industry, for instance, had one central source of control, then all healthcare vendors could sign up once and become members of a vast blockchain. That would improve efficiency," Clabby said. "Plus, everyone would be governed by the same set of rules so everyone would be on the same page. That's another efficiency."
A blockchain application stores every transaction in a distributed ledger, which can add up, particularly when it is used by multiple parties. The cryptographic algorithms used to guarantee trust also add some data overhead. These factors can lead to larger data structures that each enterprise might have to manage. Enterprises can reduce the risk of such blockchain problems by finding an appropriate balance between data stored in a decentralized ledger and other data stored in more traditional data stores used by ERP systems.
"In enterprise environments, such as a supply chain, the metadata around transactions is imperative to record, but impossible to hold in a public blockchain," said Brian Platz, co-CEO of Fluree, a blockchain database provider.
The supply chain's ERP must include context for transactions, such as a purchase order, the owner of the materials, the timestamp and the equipment, to effectively operate at scale. Early blockchain projects have operated at the business logic tier of the application stack and still need to push data and metadata related to a blockchain to a centralized, static legacy system. According to Platz, this disintegrated approach to building and scaling blockchain-powered applications is neither sustainable nor practical.
Buy-in with multiple parties
There are opportunities to improve process efficiency across enterprise boundaries. While blockchain can help on the technology side, putting it into practice requires a high level of technical and cultural coordination across all the concerned parties. Enterprises struggle with this today within a single organization.
Bennett recommends that companies launch blockchain initiatives with a sweet spot of three to 10 parties. One or two parties could possibly use other technologies to get the same benefits. Projects with more than 10 parties are likely to break down in negotiations.
"We've found that challenges are not always technology issues, but rather the definition and agreement on the new business models that blockchain usage requires," said Gil Perez, head of digital customer initiatives at SAP. "Getting everyone on the same page here can be challenging."
Companies must agree on what information to store on the blockchain, along with the format, access rights and the governance of a distributed ledger. Once that is settled, the technical question of how to store the information can become much simpler, and your chances of running into blockchain problems ultimately decrease.
"Even then, it is still a challenge to get the technical teams to agree on a common blockchain infrastructure, as there are so many different technology options and variations, which are constantly changing and evolving," Perez said.
Trust before technology
Early experiments in cryptocurrency suggested that companies might be able to replace trust with technology. However, in an operational supply chain involving physical goods and products, technology should be considered last in a formalized sequence of engagement steps, said Andrew Stevens, senior director analyst at Gartner. Enterprises should come to an agreement on governance, process and culture first.
"Managing trust is probably one of the most difficult areas to maintain and establish when dealing with a group of trading partners for which you may have a long stable history or who may be new entrants into your business operations," Stevens said.
It's important for participants to work out how they plan to concretely manage and maintain trust, and how it will be regulated once a blockchain supply chain ecosystem is implemented.
Companies cannot assume that every potential trading partner interested in becoming part of a blockchain is digitally mature enough to do so. The ability or the willingness of a company to invest in the appropriate technology systems to share and exchange data at an electronic level, as well as capture data at a physical level, should be the foundational requirements prior to any formalized blockchain discussions, Stevens said.
Integrating with ERP applications
Many of the security problems with cryptocurrencies aren't found in the blockchain itself, but rather in how they interface with an existing payment infrastructure. Similar problems are likely to emerge as enterprises tie blockchain applications into existing ERP systems and supply chain applications.
"Interfaces with existing systems present one of the biggest vulnerabilities in a DLT network, which is why it's important that APIs, microservices and containers are extremely carefully architected," Bennett said.
Application development complexity
Blockchain applications are written as smart contracts that could help streamline supply chain business processes that cross organizational boundaries.
"Many of the modeling and scripting languages available today are immature, often bug-ridden and require developers to work at a level of the stack many are no longer used to handling," Bennett said.
Although the tools are improving, this is an area that requires careful attention. Vendors are making open source efforts to address these shortcomings and developing new tools to make DLT deployment and smart contract development easier and more secure.
Updating inaccurate records
Blockchain data stores are immutable, which means it's not possible to update an inaccurate record. This inevitably makes it more difficult for a bad actor to make changes after the fact, but it also makes it more challenging to address errors caused by manual data entry, broken IoT sensors or confusion from other parties using different terms to describe the same thing.
Platz expects the industry to adopt mechanisms from accounting to correct these types of errors in supply chain data. For example, if a bank makes an error, they don't correct a bank statement, they make an adjustment. This makes it easier to identify when support agents may have used the incorrect data in their financial reports.
Bigger issues could also arise from schema changes or when data must be legally changed. These issues may require forking -- or creating a new version of the data structure -- which would require updates to the applications that use the blockchain data store.
Poor technology and use case fit
Blockchain and DLT represent a family of different approaches to creating supply chain applications with different properties around immutability, confidentiality, energy consumption and scalability. Lacking a proper understanding of how to correctly apply the technology to practical use cases will only lead to more blockchain problems.
"I believe the most promising approaches to addressing blockchain concerns start small and don't try to change the world overnight," said Monish Darda, CTO and co-founder of Icertis, an enterprise contract management platform. "Disruptions to enterprise commerce come via consensus, and blockchain is no different. Making sure permissioned and consortium blockchains are used to solve real problems where all the participants benefit, like contractual traceability before transactional traceability, will help faster and wider adoption and accrue real benefits."