, , ,

Is Blockchain Today’s Shiny New Toy or the Latest Technology for Secure, Complex Transactions?

Coauthored by Vint Cerf, Leonard Kleinrock, and Salman Baset.

When industry leaders Vint Cerf, Chief Internet Evangelist at Google, Len Kleinrock, Distinguished Professor of Computer Science at UCLA and Salman Baset, CTO Security for IBM Blockchain Solutions, get together to talk about blockchain, you know some sparks will fly. In this wide-ranging discussion, they share their views on everything from the security of blockchain to its most and least promising applications. These are their personal opinions and do not represent their organizations nor opinions of The Marconi Society.

Is blockchain today’s shiny new toy or does it have scalable application down the road?

Vint: My personal view is that blockchain is oversold and does not always scale. I don’t buy the argument that anonymous blockchain is secure. In order to be secure, you need to know who are you trusting. Some applications may work if the blockchain is built for that specific application. For example, digital signatures may work, but certificate authority does not. Cryptocurrency applications are problematic because blockchain records the transaction, but fails to protect the coin. Let’s just say, I’m skeptical.

Len: Your issues are correct, Vint, but there is a lot of work going on to mitigate these problems. There are a number of approaches to fixing trust, scale and speed under investigation, though none is proven yet at scale. The design and introduction of an effective reputation system is needed. A reputation system can address the question regarding how to evaluate whether an entity can be trusted. Some systems use networked propagation, such as page rank type of algorithms, as a crowd-sourced proxy for credibility. We can create trusted parties, but can they actually be trusted? That’s where a reputation system can provide value.

Vint: That’s right. A person cannot be distinguished from a bot and the wisdom of the crowd could be a bot.

Len: We are looking at work by Silvio Micali at MIT to relax the dependence on a small number of trusted parties (authorities) to reach consensus and authenticate transactions. Instead of asking a fixed known trusted group to authenticate a block of transactions (recognizing that the group could be corrupted), Micali’s approach is to select large random groups to reach consensus and do the authentication – no one knows who they will be before they are selected and once they authenticate for one stage of the process, they disappear before the next stage which then uses a different random set. So they are not depending on predetermined authorities. Of course, this does not address what applications to actually use blockchain for.

What are the best applications for blockchain?

Salman: First, a bit of a background on blockchain. Not all blockchains are the same. Some are anonymous and others are permissioned. Anonymous blockchains often rely on electricity-intensive consensus algorithms such as proof of work, to mitigate well-known security attacks due to anonymity, such as double spend.

In contrast, in a permissioned blockchain network, the participants are known, and thus there is less of a need to use proof-of-work style consensus algorithms. Moreover, it is pertinent to mention that nodes in permissioned blockchain networks can choose to make their ledgers public, i.e., readable by anyone with an Internet connection. The design points in both anonymous and permissioned blockchains involve tradeoffs between speed, scale and privacy.

One example of an application for blockchain is global trade, where sharing of information related to shipping containers is required as these containers traverse geographical boundaries. Another example is tracking food from farm to fork to efficiently trace food contaminations. A broader use case is settlement. Disputes can arise in settlements because settlement parties typically maintain their own view of the information. Also, in telco networks with global roaming, companies need a shared and distributed view for correct billing. Efficient global transfer of money is another important use case for blockchain.

And of course, digital currencies that may or may not be pegged to a fiat currency are an important and the most well-known use case for blockchain.

What applications and use cases can only be done with blockchain?

Salman: To identify the right applications for blockchain, we need to think about use cases, the business problem to be solved and the needs that cannot be addressed through existing solutions, such as databases. Blockchain could be the best solution for an application involving multiple parties, where parties need a shared source of truth to solve a business problem.

Let’s look at a trucking example involving perishable items such as fresh vegetables and fruits. A truck arrives at a warehouse with these perishable items. The trucking company and the warehouse each record the arrival time and the condition of perishable items, possibly through IoT devices on blockchain. That information is distributed to all parties (such as retailers) that will receive these perishable items, again through blockchain. In the case of spoilage, the tracking information can be traced to the relevant party. Given the right level of recorded information, the potential for dispute can be minimized and, in certain cases, completely eliminated.

Of course, these companies could choose to use a trusted database company as a third party to store data. In order to do that, though, each company must agree to have their related supply chain data submitted to a trusted third party. A blockchain solution is different than entrusting your data to a third party because relevant information is distributed among participants.

How does the blockchain stay up to date?

Vint: Whether it is distributed blockchain or a centrally managed database, you still need to bind information to the asset. If bad information gets introduced into the blockchain, there is a problem. So how do you bind the information to the asset so that you’re sure you have legitimate source information? For example, carriers don’t trust devices because they do not control them. Devices don’t trust the network. Can blockchain address that? And how much memory is needed if the blockchain is constantly being updated to all of the devices?

Len: This is a big issue. The information is constantly updated and we need to keep all of the copies.

Vint: All Google Docs are replicated and kept in real time, so it is possible to create real time data and keep records. The rate of change on Google Docs is relatively slow, though, and the transaction volume is not what it is in a global transaction network. There could also be a legacy issue depending on how long the data needs to be kept.

Blockchain is definitely a way of creating a ledger that is hard to fabricate and alter. It is visible publicly in the case of public blockchain and that is still the big benefit in certain scenarios. If the information going into the blockchain is false, though, it is forever part of the record.

For more on blockchain from Salman Baset, visit his blog post Securing Blockchain and the Applications of the Future.