Samuel G Brooks
Samuel G Brooks

Block8's Chief Technology Officer

25.01.2020. in Technology, Thought Leadership, Solutions, Product

Why Your API Strategy Could be a Mistake

The reliability of sharing data is crucial in designing any information system, for the speed at which you can rely upon information as correct, is the speed of business.

Most of the time, a digital transformation initiative revolves around the use or provision of APIs that access an underlying database. However, in the current context of distributed ledgers, we are provided with a potential alternative technology for reaching our business objectives.

There is a fundamental difference between distributed and centralised systems in how they achieve a data “sharing” use case (i.e. between smart contracts and APIs), and in this article we briefly compare their core data sharing mechanisms.


Comparison of Existing Technologies

Key Takeaway: “Messaging (APIs) copy information which is bad for dealing with assets or processes. Consensus systems (DLT) is a better choice in these circumstances.”

Message-based communication

Centralised systems communicate with each other using messages. This often creates the need to reconcile both databases to gain an appropriate degree of confidence before the information can be relied upon. Put simply, a business cannot make business decisions if it doesn’t know what’s true.

In a scenario where messaging is our only technology option, there is no better technology solution for intermediation than to pay an intermediary. This can be seen with stock exchanges to manage delivery vs payment use cases (DvP), such as a central market operator acting as a central counterparty (CCP). With smaller, one-off transactions, parties often engage the services of custodians to ensure the exchange of assets for consideration is not put at risk by trusting your counterparty and is reliably executed.

Disconnected datasets result in opaque value chains, meaning that insights and automatic decisions from activities up and downstream from your enterprise are impossible to create, leading to the need to engage auditing services to manually attest to the veracity of published information.

Under the hood, messaging is actually just copying data, which opens that data up to all kinds of failure modes. Copying data is generally a bad idea when dealing with things of value, but, up until 2008, there has been no technological alternative.

Consensus-based communication

A superior method to improve business communications is to use a distributed system. The reason for this is that you are using a single logical data store with your counterparties. In such a scenario, data is distributed, rather than passed from isolated system to isolated system. Systems that communicate using a consensus mechanism avoid any need to reconcile because the database that is held within any participant’s system is mathematically guaranteed to be a correct replica, with database changes applied across the entire network at a time.

While updates to a distributed ledger are slower than a single database, readout speed from the ledger is the same as any modern system.

Messaging vs Consensus

The below illustration is a direct comparison of two simplified information system stacks. On the left we have the incumbent centralised design where a user can retrieve information by sending a request that is routed through to the database at the bottom of the stack, introducing several layers of complexity where errors may arise.

On the right is an equivalent stack using a distributed system for an API strategy. The API to interface with the server is no longer needed in the same way as before because the business logic for updating the database is merged into the underlying ‘ledger’. A thin API and web server still exists in this illustration as a convenience layer for an end-user, however this is not strictly required as the blockchain network may be able to be connected to directly.

API vs distributed ledger communications


The table below provides a direct comparison between centralised and decentralised approaches for the types of information we are interested in.



(API communications )


(Smart contracts)


API required to be built to pass data between systems. 

No API needed as smart contracts define data to be shared. 


Data is copied.

APIs only offer a ‘report’ on the underlying database. Reconciliation needed for business decisions. 

Data is distributed.

Data is guaranteed to be consistent with counterparty and so no reconciliation is required.


Copied information can never attain 100% assurance it is true.

Distributed information comes with mathematical guarantees of its truth.


Immediately out of date.

Always synchronised.

Data Speed

Read: Fast  /  Write: Fast (~0.03s)

Read: Fast  /  Write: Slow (~3s)

Business Speed




Coordination needed for individual system upgrades if connected in a network.

Upgrade coordination inherited for free as all parties share the same program.


To find out more about how to best design an information system for your specific use case, check out our eBook:

New call-to-action

About the Author: 


Samuel G Brooks is Block8's Chief Technology Officer. He is an expert in developing decentralised software products and has designed numerous solutions for startups, enterprise, government and OpenTech since 2014.

Samuel regularly speaks at technology conferences, meetups and podcasts, and holds several advisory positions on technical industry boards and committees. He is a also a heavy contributor to blockchain and fintech-related public inquiry and writes about the nature and benefits of distributed ledger technology on our blog.

Samuel holds a degree in Electrical Engineering from UNSW, and has stayed close to both the code and the latest research ever since encountering Bitcoin in 2011.


Subscribe to our Newsletter

Get the latest news on our products or learn what's happening in our guilds.