The 451 Take
If you want immutable records chronologically linked together in 'blocks' and you're part of a trustless 'chain,' you're going to love blockchain. If not, the chances are that a more traditional distributed database or some combination of a distributed database and blockchain-like features is likely to fit your needs. Additionally, note that the effort required to run a blockchain application or join an existing blockchain network is considerably less than if you're intending to develop one yourself. Do not succumb to the hype; be practical instead. Do a careful analysis of your requirements and how those are met by different database technologies, including distributed ledgers and blockchain.
Trustless chains
Blockchain has an advantage over traditional centralized databases for the use cases described above. However, not all traditional databases are centralized. Some distributed databases might be suitable, architecturally, for a distributed network of participants – particularly compared with permissioned blockchain networks, and especially those where the participants are members of a single entity, where established trust is a prerequisite. In effect, the more permissioned and trusted the environment, the more likely it is that a 'regular' distributed database system can do the job.
To delete or not to delete...
Blockchains are good for use cases that need an audit trail, a continuous list of immutable records (transactional data – e.g., user activities and interactions) that cannot be deleted. Blockchains can provide a chain of custody and a single source of truth for the transfer of any 'digitized' asset, be it money, diamond, contract, land, intellectual property, identity or other. When a record is accessed or altered, the activity is logged into and verified by the blockchain.
While blockchain can be used to create and read immutable records, if you need to update or delete data, then blockchain is not suitable for storing the data itself. In other words, blockchain only addresses the first two of the CRUD (create, read, update and delete) functions of persistent storage.
As a result, in relation to the EU's new General Data Protection Regulation (GDPR), it is not clear whether and how anyone can store personal data on a blockchain and comply with the literal or absolute interpretation of the 'erasure' right/obligation stipulated by the GDPR. It further complicates the issue that the GDPR defines 'personal data' and 'identifiability of a natural person' very broadly. For example, is a natural person whose encrypted/hashed identity data is recorded on a blockchain identifiable? One potential workaround is to store all personal data off-chain and only record transactional data (e.g., user activities and interactions) on-chain. However, there are digital identity offerings with the promise to store encrypted identity data across multiple nodes on the blockchain, where only the private key holder can decrypt the data. At the very least, businesses need to be aware of the risk in operating blockchains that include personal data of EU-based individuals.
On a more general level, for regulatory and compliance reasons, if you want to use blockchain, you will have to conduct thorough research, possibly do some adjustments and work with your regulator.
In terms of potential cost savings, if you use a running blockchain application or join a blockchain network, you can rely on other organizations when it comes to support and maintenance. However, if you are developing a blockchain-based application yourself and need to provide support and maintenance for it, it has to be assessed whether that's less costly than, for example, having in-house administrators keeping a distributed database running.
Performance...penalty?
In a permissionless blockchain, every node needs to process and validate every transaction. For the network to get faster, more compute would need to be added to every (public) node, which is something that is hard, if not impossible, to control. The type of consensus protocol used affects the performance of the network. Proof of work (PoW), in particular, unfavorably affects transaction-processing speed. There are a few approaches, such as the Lightning Network and the Raiden Network, that are attempting to improve the performance issues of public blockchains such as Bitcoin and Ethereum. For example, the Lightning Network is an upgrade proposal that allows peers to quickly and inexpensively transact among themselves on private channels, with the final outcomes broadcast to the Bitcoin blockchain. Startups and consortia are tweaking existing protocols, as well as developing and experimenting with new algorithms, to tackle the flaws of the PoW and proof-of-stake (PoS) consensus protocols. Delegated PoS is an emerging, reportedly highly performant, algorithm used by BitShares, the
In a
Akamai and Mitsubishi UFJ Financial Group recently announced their plans to launch a new blockchain-based online payment network designed to execute more than one million transactions per second (with latencies of less than two seconds per transaction). Akamai claims that the underlying
The idea behind decentralized data marketplaces/exchanges – an emerging concept that leverages blockchain technology for buying and selling personal and business data – is to equalize the access to data and unlock its value for individuals, as well as researchers and businesses (again, set it free, but not for free). As mentioned above, however, beware of limitations in terms of a provider's ability to update and delete data. The intention is that data owners/originators – whether individuals or organizations – are able to easily and securely share and monetize their data on their own terms, which is a concept that we believe will be welcomed by many, but what will actually work in the real world is still unclear at this point.
If fault tolerance is important to you, we believe that blockchains have a decisive advantage over more traditional distributed databases. Decentralized systems like blockchains are less likely to fail accidentally and more expensive to attack, and it is a lot harder for participants to collude to act badly (benefit themselves at the expense of others). Getting the same level of robustness in a 'regular' distributed database – especially for use cases that involve multiple parties – would be extremely difficult or unreasonably expensive, if at all possible.
How big are your data files?
Blockchains are generally not well-suited to store large data files, at least not today. Blockchains typically store
451 Research Blockchain Taxonomy
Understanding the taxonomy of blockchain and related technologies is the first step in making informed decisions. Below is our initial taxonomy, which will be updated as the market evolves.
Csilla Zsigri is a Senior Analyst for 451 Research’s Cloud Transformation channel. Csilla also covers blockchain and works on custom research, providing strategic guidance, as well as market and competitive intelligence, to technology vendors, service providers
Matt Aslett is a Research Director for the Data Platforms and Analytics Channel at 451 Research. Matt has overall responsibility for the data platforms and analytics research coverage, which includes operational and analytic databases, Hadoop, grid/cache, stream processing, search-based data platforms, data integration, data quality, data management, analytics, machine learning and advanced analytics.
Keith Dawson is a principal analyst in 451 Research's Customer Experience & Commerce practice, primarily covering marketing technology. Keith has been covering the intersection of communications and enterprise software for 25 years, mainly looking at how to influence and optimize the customer experience.