Enterprise Blockchains and the Data Integration Challenge
If you thought that blockchain technology started with Bitcoin, the most famous of all cryptocurrencies, with publication of the white paper by Satoshi Nakamoto in 2008, you would be missing an interesting piece of history. Blockchain, the underlying technology of Bitcoin, had its origins
nearly two decades prior with two physicists trying to find a solution in knowing for sure about the past in a world with digital data (The Eureka Moment that Made Bitcoin Possible: https://www.wsj.com/articles/the-eureka-moment-that-made-bitcoin-possible-1527268025 ). The fundamental issue to them was about protecting what we know about the past.
Consortium and Private Blockchains
Fast forward almost 30 years, a key value proposition of Blockchain remains its immutability that lends to the ability to track changes. This is of course key to operating in a trustless environment. But it need not be limited to a 100% trustless public environment, such as Bitcoin, but useful in permissioned and private blockchains as well. There are numerous postings on the differences and benefits of the different types of blockchains, with one of the earliest ones I saw from founder of Ethereum – https://blog.ethereum.org/2015/08/07/on-public-and-private-blockchains/
When we talk about enterprise use of blockchain, we would usually be concerned with permissioned and private blockchains. A permissioned or consortium blockchain is where membership is controlled by the consortium and the consensus process is also governed by the consortium’s own rules. Read/write permissions to the blockchain are also controlled such that members may have read/write access, while regulators or 3rd parties could have read access only. Energy Web https://energyweb.org/ is an example of a consortium blockchain for the energy sector.
Fully private blockchains, on the other hand, are private to an organization and therefore easier to manage compared to the consortium blockchains. Much like the consortium blockchains, rules are set by the organization, with read permissions could be given selectively to the external entities, such as regulators. Private blockchains may be the ideal starting point for enterprises. There is likely to be fewer number of nodes, so reaching consensus is easier, which will lead to faster transactions. One of the key challenges of blockchains is scalability, especially in comparison to enterprise transaction velocities, which are relatively much higher today.
The Data Integration Challenge
The next question may be why not a centralized database instead of a private blockchain. The answer lies in what the two physicists who had the eureka moment that led to blockchain: we protect what we know about the past. Even with private blockchains there is more trust with disintermediation among the entities in an organization at the cost of loss of some privacy (Gideon Greenspan – Four genuine blockchain use cases – http://www.the-blockchain.com/2016/06/05/gideon-greenspan-four-genuine-blockchain-use-cases/ ). In a centralized database, there is always the risk that a user with sufficient privileges (e.g., DBA, SSO) can overwrite and corrupt the contents of a database.
But while enterprises adopt blockchains where disintermediation is important, traditional databases will continue to play a critical role in transaction processing and analytics. In that regard, the challenge is how we integrate these stores of data. There are several reasons why traditional databases and blockchains will co-exist:
- Limited storage of data on blockchains – blockchains store the hashes of data, but there may be contextual data that needs to be stored off-chain
- Real-time analysis of blockchain and business data – enterprises will need to analyze business data along with associated transactions held in blockchains
- Centralized applications written for a relational database with a SQL interface must continue to be available as new decentralized applications get developed on blockchain
(From https://blog.iota.org )
For example, consider an auto manufacturer that is in the ride sharing business. The manufacturing operations already use Enterprise Resource Planning (ERP) solutions that are supported by enterprise-strength relational database systems. The inventory of vehicles manufactured is entered into the ERP system with various attributes including VIN, color, vehicle type, mileage, cost of recalls, etc. The ride-sharing business uses a blockchain platform to record each ride, including mileage, service costs, etc. The auto manufacturer could be interested in integrating this data and doing analytics to understand the lifetime costs per mile for a specific vehicle, for example. This requires data integration solution to transfer data between the disparate systems.
Who best to enable this blockchain-fired sharing economy than SAP with its solutions in business applications, enterprise data management and blockchain services?