Blockchain Works With Systems of Record to Integrate Data
Distributed ledger technology blockchain links with existing systems of record (SOR) application servers and data managers. They share information and assets securely and immutably between external parties.
It can be used to track the lifecycle of a car, a financial transaction or sources of food ingredients. (Learn more about how various industries are using the technology in “The Business of Blockchain”.) The transactions are very different from those performed on mainframe transactional systems in terms of goals, timing and performance.
Simply stated, a blockchain transaction is a contract on blockchain, says Donna Dillenberger, IBM Fellow. The goal is to share data across multiple enterprises in an auditable, immutable way.
Blockchains today don’t provide millisecond response times, and a single blockchain network won’t support billions of transactions a day. However, a blockchain transaction will keep all parties on blockchain informed and will provide a record for all parties. The transaction is shared across enterprises, so many views of the immutable data are given.
In contrast, SOR transactional systems on z Systems* use transactional managers such as IMS*, CICS* and DB2*. Transactional managers process billions of transactions a day at millisecond speed with high volume, high throughput and high qualities of service, Dillenberger notes. They ensure each transaction has four attributes: atomicity, consistency, isolation and durability. When transactions are committed across all copies of that cluster of transactional managers, one view of the data exists or it doesn’t get committed at all, she says. Typically, traditional SORs don’t support multiple live copies in the manner blockchain does—although some specialist deployments of DB2 use replication for high availability.
Blockchain Uses
Enterprises are looking to use blockchain with existing SOR data managers for two main reasons:
1. An enterprise wants to share SOR data with external parties and leave a record of that data exchange. “The record would include a timestamp and list of what data was shared so the record can be audited and the data is immutable,” Dillenberger notes.
2. An event on the blockchain generates a need to interact with a participant’s SOR. That blockchain event may trigger the start of a CICS transaction or a DB2 procedure.
No matter why blockchain is deployed, it tracks the lifecycle of an asset whether it is a car, a financial asset or a real estate title, says Ian J. Mitchell, Distinguished Engineer, z Systems software. Blockchain offers transparent management of that asset between participants and requires tracking the asset inside the participants’ existing SORs.
Blockchain transactions might be initiated by a participant’s SOR because of an event in the asset’s lifecycle. After that change is published via blockchain, other participants can note the change and reflect it in the SOR.
The replicated nature of blockchain provides data resilience. All parties will need to agree on what will be managed before creating the blockchain contract and then create the actions allowable under the contract.
“Uniquely with blockchain, all the participants see and agree on those allowable actions,” Mitchell says.
Further, strong security credentials are required for interacting with blockchain, which means having a good registry of security credentials is important, he notes.
The Role of APIs
When blockchain interacts with the SOR, it will be interested only in the data pertaining to the asset, not all of the data that exists in the database. Blockchain uses APIs to identify the data it needs.
Blockchain data is posted in logically separate systems, and APIs are used to interface with SOR, Mitchell says. IBM is enabling Representational State Transfer (REST) APIs to work with many of its existing offerings such as z/OS* Connect. REST APIs identify certain resources and use HTTP to access them. Systems integration on blockchain with DB2 and CICS are done with REST API.
A blockchain framework tool called Fabric Composer, which was developed by IBM with the open-source Hyperledger community, creates REST APIs for the assets modeled and managed in the blockchain. z/OS Connect can use those REST APIs to tell the SOR to invoke actions in the blockchain. “z/OS Connect makes the interaction between the blockchain and the existing SOR as consistent as other services, which react with the SOR,” Mitchell explains.
REST APIs also will work from the blockchain to the SOR. IBM is working on initiatives that incorporate the bidirectional functionality of REST and APIs.
It is also supporting SOR integration initiatives to provide APIs that can be invoked from the SOR and that can be used by blockchain to access the SOR. “Bidirectional APIs are the best way for SOR to be driven by other services,” he notes.
IBM’s Solution
IBM launched IBM Blockchain for Hyperledger Fabric V1.0, which enables a host of data stores to be used with blockchain. General availability is slated for late spring-early summer.
IBM Blockchain works with a JSON data store like CouchDB, a key value data store like LevelDB and with relational databases like DB2.
“Using IBM Blockchain, you can run analytics directly off of DB2 or CouchDB without having to offload or export all of your data,” Dillenberger says. A DB2 table could be configured as an Apache Spark resilient data set so you could apply Spark machine-learning models to both the SOR and the blockchain tables. With other blockchains, data must be exported into the technology before it can be used.
This analytics capability can be very valuable. For instance, a car can be tracked on the blockchain from its manufacturer through various owners to the end of its life. A SOR DB2 table at a car dealer may have such information as who owned the car before the current purchaser. The car dealer may want to run analytics against all of the owners to check on maintenance and other factors. Even though some of the data isn’t on blockchain, it would be in the SOR DB2. The SOR and blockchain data would be viewed as one data lake by the analytics.
What’s Next?
IBM Blockchain is just one of the advances that IBM is rolling out for blockchain, according to Dillenberger. To facilitate the high-security business network that IBM provides for production clients using blockchain, the company is increasing its cloud capacity in Asia, Europe and the Americas. IBM also is providing additional security services for blockchain peers using isolated Docker containers.
IBM Fabric Composer helps with SOR integration by enabling the customer to model classes of assets on the blockchain.
“You’ll be able to say the original source for this data class is SOR and this blockchain framework tool will generate the APIs necessary to get that data from the SOR,” Dillenberger confirms.
Further, IBM is providing analytics services to allow blockchain clients to find trends, anomalies and patterns in their blockchain data.
Blockchain is evolving to meet real world needs. This is very much an open-source-driven evolution, which IBM is supporting.
“Enterprises need to become involved with the open-source blockchain community because that’s the way enterprise-level requirements are included in the solution that IBM provides,” Mitchell says.