A group of people surrounding Ken Thompson and Dennis Ritchie were quite busy at Bell Laboratories shaping the world as we know it today. They invented a milestone in programming languages — C, the breakthrough operational system Unix, the first shell, UTF-8 and a long list of others that alone would have been enough to receive a Turing award.
How were they so productive? Of course, the density of talent at Bell Labs was as high as within the Beatles (before John married) — but one tiny detail helped drive them. And that can be summed up in one word: focus; finding out what holds the application together in its innermost folds and excelling at that. It takes writing programs that do one thing and do it well. Today, this set of guiding principles is popularly known as the unix philosophy.
It’s loosely coupled with another imperative guideline governing software architecture: “Choose the right tool for the job,” with the right tool defined as one that focuses on solving a problem that is as well-suited as possible to the challenge at hand. With this principle in mind, we set out to find the ideal blockchain to address the problems we want to solve at DSTOQ.
We want to bring real-world assets into the world of crypto and require secure settlement protocols with a special focus on tokenizing assets. Stellar is a settlement layer for arbitrary assets. It does this one thing and does it incredibly well. It turns out that this is in-sync with much of our business requirements.
The focus on settlement releases Stellar from the burdens of a Turing complete smart contracting language. Although Stellar supports operation-chaining through transactions, very complex logic is excluded by design in order to guarantee secure settlement.
However, representation of real-world assets requires some additional properties that other blockchain solutions do not ship as a first class citizen — simply because they focus on different use cases. These properties are a prioritization of safety over liveness and asymptotic security.
Each consensus algorithm has to decide on two of three features: fault-tolerance, safety or liveness; having all three is not possible; this result is mathematically proven. Most protocols choose fault-tolerance (the network can still progress even if some nodes fail), but they do vary on liveness vs safety. For example, Bitcoin opts for liveness — if consensus cannot be reached, the network may fork , prioritizing progress over unambiguousness. This happened in the past, leading to multiple communities claiming to be the real Bitcoin.
This is unacceptable for DSTOQ because of the legal implications of an asset that are tied to a real world value but are represented multiple times on different ledgers. The Stellar Consensus Protocol (SCP) favors safety over liveness, meaning that the network, faced with the same circumstances, comes to halt until consensus can be reached. A side effect of this is low latency, as there is no need to wait for multiple confirmations to be sure that a transaction is included on the primary ledger.
Asymptotic security is a feature that builds upon a function that can be tuned in a way so that even the universe may not be big enough to compute it. To continue with the Bitcoin comparison, proof of work does not exhibit asymptotic security in the sense that everyone can join or leave the network, with the contribution to consensus being measured by a provisioned hash rate. Therefore, a network takeover can be performed by controlling more than 50% of the current hash rate. The computing power required to do so exists by definition.
SCP builds upon a map of trust between validators named the quorum slice. Since an attacker cannot simply spin up nodes as they must also gain trust to this interconnected map of validators, the only possibility for them is to fake validations of existing nodes , which are signed by an elliptic-curve algorithm that cannot be cracked with existing computing powers. This algorithm can theoretically be tuned forever, hence asymptotic.
The reason Bitcoin is safe is that the resources required to pull of a 51% attack exceed the gains for two reasons: 1) The computing power required is huge 2) the value of your stolen goods would crash once the attack goes viral.
However, since DSTOQ digitalizes real world assets, one has to take into account these tokenized assets. On a blockchain that is securing by hashrate, it could actually be worthwhile for someone to do the 51% attack even though it is expensive in the native currency terms because they are speculating on capturing the real-world assets. This attack vector is neutralized in asymptotic security, as featured by the Stellar Consensus Protocol.
The amazing thing about the above-described features is that they were not added in order to bloat the product with more use cases. They were added in order to effectively carry out asset tokenization.
So we think we chose the right tool for the job — to bring real world assets on chain, enabling everyone to participate in economic growth. We want to rewrite the global investment rulebook and we are excited to do it on top of Stellar.
About Christian Peters:
Christian is co-founder and CTO at DSTOQ. He was previously working in blockchain consulting at Chainwise Group and as a freelance software architect for all the big names in the Berlin startup scene.