The blockchain has caught the fancy of banking executives, who see the system as a way to overhaul antiquated IT infrastructure. To be sure, the technology underlying much of finance is outdated by today's standards. With startups from Silicon Valley at the gate, banks are rightly looking to differentiate themselves and win customers through better technology. Blockchain provides the best technology available today for many applications, especially in the back office.

Blockchain technology has several appealing features for the financial industry. Different parties can easily create a consensus on the current state of events, without having to transfer proprietary data back and forth among firms to be validated by each party. Add in smart contracts, and a lot (not all) of simple rules can easily get automated in the process. It is no wonder that many banks are creating internal working groups and consortiums to experiment with how the blockchain might fit into different markets.

However, looking at blockchain as an application that solves some immediate business need may turn out to be a mistake in the long-run evolution of this technology. It is more plausible that the technology will evolve in a way that makes the blockchain a protocol layer, on which individual applications are then built. Such a system would be preferable in a global marketplace where no financial institution has to play in a system built by a competitor. It would also be preferred by regulators, because it standardizes applications and smart contracts, thus making it easier to monitor systemic risks.

The blockchain does one thing really well: help different parties come to an irreversible consensus. That should be the protocol layer. Various applications can then be built on top of that foundation. For example, it is entirely possible to build an application that provides reversible transactions on a blockchain. Or create an automated piece of code (a smart contract) that "pays a dividend" if certain conditions are met.

Given the base protocol blockchain layer, the industry can standardize and innovate on two aspects: defining financial instruments and defining processes around them.

With blockchain as a protocol, the industry can come to a consensus – based on defined rules – on what constitutes financial instruments, from simple stocks and bonds to more complex swap options and collateralized debt obligations. The market will know an instrument when it sees one on the blockchain.

In terms of defining the processes, blockchain as protocol can establish the characteristics of smart contracts, from settlement and custody to repo agreements and voting, that are carried out on the protocol. When does a bond pay a coupon, and how does it reach the beneficial owners? When is a swap option exercised? What happens when a firm defaults, holding on to repo agreements? When all such processes are built on top of a single protocol, it is easier to get a holistic picture of assets and systemic risk in the system.

The alternative is that each pocket of the financial industry – CDOs in North America, commodities in Frankfurt and futures in China – build their own blockchain applications and corresponding smart contracts on their system. Such applications are unlikely to integrate well with other blockchains. It also makes it unlikely to be adopted by competing firms or consortiums. It makes the job of the regulators harder, since they need to vet many versions of blockchains and smart contracts promising to do the same or similar things.

It is always possible to build restricted applications on top of open protocols (similar to online banking, which is a restricted application, built on top of the open internet protocols). If the underlying blockchain is a global protocol that financial companies from across the globe can access, the applications built on top of it can still restrict access and usage according to local laws. However, it would allow startups and smaller firms to innovate at the same level as large, global entities.

An approach of building out a base-layer blockchain protocol with innovative applications on top of it would take more time to come to market than individual applications. However, over the long term, such an approach would provide better means of innovation and level the playing field among the industry participants. It would also help standardize a lot of the industry, thus making the market for financial services more transparent and cost-competitive.

Siddharth Kalla is the chief technology officer of Acupay, a technology provider in New York specializing in cross-border finance. The views expressed are his own.