How to integrate blockchain with legacy systems (and whether you should)

Blockchain is not middleware meant to tie into existing legacy systems, but there are methods for automating the flow of data from ERP systems to a distributed ledger technology. First, though, you should figure out if you even need it.

binary chain / linked data / security / blockchain
Ismagilov / Getty Images

Blockchain offers enterprises a new method for transacting over a distributed, trusted network, but plugging the technology into existing databases, ERP systems and a client/partner base is no small task. And in many cases, it isn't even needed.

While few production blockchains have been deployed, the distributed ledger technology (DLT) was still one of the most-hyped technologies of 2018. It's not just hype; the blockchain market is expected to skyrocket in value from $708 million in 2017 to $60 billion by 2024.

Because of all the market hype, companies made a mad dash to implement DLT, lest they lose a competitive edge, according to Kevin McMahon, director of emerging technologies at Chicago-based consultancy SPR.

For many companies, however, DLT isn't a good fit for tasks that can be just as easily handled with traditional technologies, such as relational databases. For others, the challenges associated with implementing DLT will have less to do with the technology itself and more to do with building out a network of users who can agree on governance rules.

"The technology part isn't really all that difficult. It's novel, the cryptography's great and it's got some cool features, but the real challenge is building out that network –  finding people who want to participate and want to share data amongst themselves and are committed to maintaining the infrastructure necessary," McMahon said. "It's about making sure their processes and workflows are able to accommodate writing additional data to a blockchain."

blockchain survey infographic SAP

Brigid McDermott, vice president of the DLT-based IBM Food Trust, said blockchain as a food supply chain management system addresses two major business needs: produce provenance and the need for a single standard for interoperability.

McDermott compared blockchain to the Betamax and VCR videotape format wars of the late 1970s, when the better technology – Betamax – lost out because VCR tapes had more industry support.

"The VCR [industry] built content," McDermott said. "Blockchain is very similar. It doesn't matter how good a technology is if you have nothing to use with it."

IBM is attempting to bootstrap its blockchain cloud service through proofs of concept  such as Food Trust, a produce supply chain management system and TradeLens, an international cargo tracking system.

Maersk has been testing TradeLens with 94 partner participants and Walmart is piloting Food Trust, even going as far as to tell its produce suppliers to join the network by September of this year.

In one pilot of IBM's blockchain, Walmart tracked where a mango had been, what store branch it came from, what packing house it went through, in what cold storage facility it was kept, and what distribution center it passed through.

Much of the information about the mango shipments came directly from Walmart's ERP system, McDermott said, eliminating the need to create a secondary record or re-create shipment information manually. Walmart's traditional supply chain tracking system took three weeks to trace the origin of mangos; IBM's Food Trust blockchain dropped that time to two seconds, McDermott said.

Walmart blockchain Walmart, MIT, IDG

Frank Yiannas, formerly Walmart's vice president in charge of food safety, and current deputy commissioner of Food Policy & Response at the FDA, explains how blockchain reduced the time to track the origin of a package of mangoes from a week to 2.2 seconds.

Blockchain, however, doesn't integrate directly with ERP systems, spreadsheets or databases. Instead, APIs and data-sharing standards, such as GS1 (best known for the machine-readable barcode protocol), have been used to enable interoperability with legacy data systems.

The IBM Food Trust, for example, avoids manual data input by leveraging legacy tech investments through the GS1 standard; it automates the transfer and understanding of data between different parties on the electronic ledger.

"We translate GS1 standards into files that can be easily ingested into our solution via API," McDermott said. "We've been figuring out what connectors will optimize usage of a legacy system – how to connect an existing ERP system into our blockchain data. Some businesses use Excel spreadsheets, but the idea's the same."

Regardless of how blockchain is implemented, the lion's share of the cost and legwork for rolling it out is garnering business partner participation in the network and involves hammering out business agreements and governance rules, McMahon said.

Standing up blockchain nodes (servers) or purchasing blockchain-as-a-service from IBM, Oracle, Microsoft or SAP is really no different from running cloud business applications from Amazon Web Services, Microsoft Azure or Google Cloud.

"Putting together the governance model and putting in the effort, time and energy building out a consortium as well as solving business challenges – that's always been the surprise for our clients," McMahon said.

Additionally, enterprises must consider that potential blockchain network partners may be wary of sharing data, and concerned over not being able to exert any leverage over network governance and rules.

IBM, blockchain, Maersk Maersk

Ninety percent of goods in global trade are carried by the ocean shipping industry each year. A blockchain pilot from IBM and Maersk includes 96 partners tracking cargo shipments.

Blockchain is best for companies who work in a network of separate but interconnected actors, like those in an extensive supply chain, McMahon said.

For most companies looking to streamline internal IT, productivity or operational issues, blockchain is a costly and time-consuming option that sometimes does exactly what other traditional – and less costly – solutions already provide, McMahan added.

"Blockchain is going to allow you to integrate and share data in a reliable manner that's tamper resistant, but the use case that makes a lot of sense in an enterprise context...is supply chain, logistics and being able to track provenance, like where did this head of lettuce come from or what's the history of this asset as it's been traded," McMahon said. "But a lot of use cases don't fall into that sweet spot."

"We've done some PoCs with clients that do have a pretty good use case, but a lot of times we spend a lot of the initial conversation with them explaining some of the drawbacks of blockchain," McMahon said.

The value proposition for blockchain boils down to two things:

  1. It can validate the origin of data, (i.e, at this time this entity input this data on an immutable ledger).
  2. It enables the use of smart contracts, which executes any business process logic implemented across the network.

Martha Bennett, a principal analyst for Forrester Research, agreed that distributed ledgers are a team sport.

"It's about data you can trust to the highest degree possible and it's about sharing," Bennett said during Forrester's New Tech & Innovation Conference last July. Bennett said public and private organizations must first determine what business processes DLT can address – and those to which it cannot be applied.

To determine if there is a use case, enterprises must first understand while many existing technologies, such as relational databases, can already address most transactional business needs, they cannot match blockchain's key attribute: collaboration.

Bennett offered up a "checklist" for companies to consider when determining whether to deploy blockchain, including:

  • When multiple parties need the same data and the ability to write to the same data store;
  • When all parties need an assurance that data is valid, and hasn't been tampered with;
  • When a current system is error-prone, too complex, too unreliable or full of friction points;
  • When there are good reasons not to have a single centralized system, such as a relational database.

IBM also suggests determining first whether a company will be able to take advantage of blockchain's attributes and whether it will scale to meet needs. Also consider regulatory risk, skills availability, and technology support. Like any enterprise application, there needs to be an appropriate level of support for the entire stack.

Last year, IBM published The Founder's Handbook, a best practices guide to implementing blockchain, which emphasizes limited use cases.

"One of the epiphanies people have when they're evaluating blockchain is at its core you could say it's a glorified database and that's one way to look at it, but it's an integration technology in and of itself," McMahon said, referring to its ability to enable collaboration among business partners. "It's not middleware used to glue or piece together existing systems. When you realize that, you realize it's not something you deeply integrate."

Copyright © 2019 IDG Communications, Inc.

Bing’s AI chatbot came to work for me. I had to fire it.
Shop Tech Products at Amazon