According Nazarov, some DeFi builders either don’t take data quality problems seriously or they underestimate the knowledge required to create and maintain such systems.
As such, it is obvious the careful measures have to be implemented in how blockchain applications get their data. Deficiencies in these areas have led to some of the latest attacks in DeFi platforms like MakerDAO and bZx.
The Dangers of Underestimating Data Quality
Nazarov believes that the problem with most DeFi builders is that they “ignore data quality”, as well as the “quality of their node operators”.
This, he believes, stems from the fact that most DeFi companies lack awareness in how data mining works. Major finance data companies like Bloomberg, Routers, and others actually invest a lot of time and money in smoothing out data-related risks. And replicating this is a lot easier said than done.
Yet allegedly some DeFi companies completely ignore these risks and build an oracle mechanism and data aggregation system as a result. As a result, they make all the rookie mistakes that big data companies don’t have to. This is the main reason why Chainlink avoids this route.
According to Nazarov, underestimating data-related risks is “very very dangerous”. In that sense, they can be viewed as ticking time bombs. And he speculates that the only reason these systems have yet to blow up as much he fears is that:
- the numbers aren’t high enough yet
- there are other vulnerabilities that are currently higher on the cybercriminals’ target list
He also points out that some entities might have already begun or are in the process of exploiting these platforms. Regardless, this issue should be addressed as quickly as possible.
How Chainlink Manages Data
What sets Chainlink apart from other oracle platforms is that they leave data collection and quality control to the experts.
Instead of solving the data quality problem themselves, they focus on building a secure and reliable oracle mechanism.
Moreover, they intend to work with as many reliable data aggregators as possible with the aim of minimizing market coverage risks, manipulation attacks, and a whole lot of other unforeseeable attacks.