Public vs. Private Blockchains – Which One’s Right for You?
While blockchain is currently one of the most trumpeted technologies, some dismiss it as a fad related to cryptocurrency. The reality, however, is that companies across industries are looking at how blockchain can address major business issues, particularly with respect to speeding supply chain and reducing friction. Ignoring enterprise applications of this technology could damage your organization’s competitive position.
The blockchain concept is not new. In fact, it’s a repackaging of existing technology constructs. It was first described as a chain of blocks that can be implemented using an electronic ledger. Once something is written on that ledger, it cannot be rewritten or modified, and is thus immutable.
When evaluating the technology it’s important to understand the difference between public and private blockchains. For an intra-enterprise implementation, adopting a public blockchain approach is unnecessary and inefficient. Public blockchain models consist of complex algorithms that are expensive, time consuming, are not beneficial to an individual business, nor required.
Public blockchains, which are best suited for inter-enterprise strategies, are often criticized for slow performance, mostly due to use of “trust no one” consensus algorithms. These are algorithms everyone or every entity has accepted as the gatekeeper for allowing data into the blockchain; because multiple enterprises are involved, “trust no one” approach, incurring the overhead of proof-of-work/time/stake validation models, is more appropriate.
But within a single enterprise, the rules for private blockchain can be simpler because trust is already defined through user ID’s, physical security tokens, biometrics and passwords. In the case of machine data, consensus could be a private identifier embedded in the machine and a security token provided to the device. If the two items pass authentication when data is sent to be placed on a blockchain, they are allowed, otherwise the request would be denied.
Beyond the ledger and consensus algorithms, another important piece of a blockchain is the use of smart contracts, which are microservices with a trigger, perhaps a Function-as-a-Service (FaaS) defining when they should execute. They must be fully audited, proving that they were executed, including the state of the system when it was executed.
Those evaluating blockchain stacks often discover limitations early in the process. The limitation most often seen is based on antiquated, heavily decoupled, models that require a substantial amount of ETL and data movement steps, leading to system latency and poor traceability. Additional complexities hide in these systems until either requirements for analytics are brought to bear, or upon a request to show the state of the system in a proven way at a point-in-time.
For most organizations, storing data in a blockchain is a paradigm shift, arguably in a good way. Choosing to adopt an event streaming system capable of storing the complete history of all the data, along with support for omni-directional global replication in real-time, is the foundational requirement. Then add on the core requirement for the ability to perform analytics in-situ.
In many blockchains, performing analytics over data requires a sidecar database, but that is a side-effect of a questionable architecture. A proper blockchain technology (i.e. event stream-based) can support analytics without moving the data, and should be adopted up-front because validating the data in the blockchain is a necessity. The moment it is moved to another location (e.g. sidecar database) that does not have immutability, the rules of governance and proof have been broken. If the system “requires” the data be placed in a secondary storage location to perform analytics then the entire value proposition of the blockchain to create irrevocable proof is destroyed.
Adopting blockchain within an enterprise can drive significant benefits. However, a private blockchain can deliver these benefits with a purpose-built approach without the cost and performance overhead of a public blockchain. Don’t dismiss blockchain out of hand. Focus on how a distributed, immutable ledger with integrated analytics can reduce the friction, time and resources that impact the delivery of your organization’s products and services.
Jim Scott is director of enterprise strategy and architecture at MapR Technologies.