What if you could outsource any computing task to the crowd without risking an epic fail? A newly developed set of technological tools could make this a reality.
With more than 11 billion devices connected to the internet, the potential power of computer networks is almost beyond comprehension. But in actuality, networked solutions are only as powerful as the trust they inspire.
Uber, for example, disrupted the taxi industry by creating brand value that conquered consumers’ qualms about climbing into a stranger’s personal vehicle. One of Amazon’s advantages in becoming the world’s largest cloud computing company was its track record in e-commerce—who better to trust with your data storage needs than “the everything store”?
Therefore, the evolving model known as Web 3.0— in theory, a fully decentralized system that could tap the mind-boggling computing capacity of those 11-billion-plus devices—has a serious trust deficit to overcome.
Jiasun Li, an associate professor of finance in the Donald G. Costello College of Business at George Mason University, is actively exploring leading-edge tools for “efficiently organizing our digital future” within the Web 3.0 ecosystem.
As a member of the inaugural cohort of Public-Private-Partnership (P3) Faculty Fellows, a program offered through Mason’s Institute of Digital InnovAtion (IDIA), Li has received a $75,000 grant for research focused on verifiable computation—a domain that concerns easing the burden in authenticating tasks within the cloud.
For example, a company could outsource the computation-intensive training of AI models to a decentralized network of servers, secure in the knowledge that the task was carried out as specified with much lighter computational overhead.
Theoretical research about verifiable computation dates back several decades. But the algorithmic advances required to make verifiable computation possible at scale are fairly recent innovations.
“Examples include a variety of Succinct Non-Interactive Argument of Knowledge (SNARK) protocols such as Plonk and Groth16 as well as a set of proof-of-storage protocols,” Li says. “Apparently, verifiable computation appears suitable for mitigating trust concerns and fully unleashing economic efficiencies of our digital future, so we will evaluate the extent to which this idea applies.”
For this research project, Li is partnering with Protocol Labs, a pioneer and leader in decentralized storage technology. In 2017, Protocol Labs introduced the Filecoin system, which enables independent providers to rent server space in exchange for cryptocurrency.
“The partnership with Protocol Labs gives me access to their talent to thoroughly understand the technical details and design choices of Filecoin. It also allows me to receive immediate feedback as my research progresses,” says Li.
Based on this collaboration, Li is currently building analytical models to evaluate the relative costs and benefits of traditional (i.e. on-premises), cloud-based and fully decentralized storage solutions. In particular, Li is looking to discover whether and when virtual computation produces aggregate gains for consumers, over and above centralized models. If the inescapable trust problem can be resolved only by an intermediary with significant market powers, this may limit or complicate the efficiencies delivered by the system. As Li points out, the purported economic upsides of cloud-based systems may merit scrutiny, in light of the well-documented relationship between monopoly and scarcity. “It is [equally] debatable whether the penetration of ride-hailing services into the turf of traditional taxi services has necessarily resulted in lower riding costs in the long run,” says Li.
Li chose data storage as the context for this seed project (which runs through mid-2025) because of his established ties with Protocol Labs, as well as the storage sector’s essential role as “an indispensable foundation of all computational tasks”. His goal for the P3 fellowship is to produce a proof of concept that will attract funding from federal agencies such as the National Science Foundation and/or the private sector for the development of special-purpose protocols applicable to storage as well as other digital services.
- October 22, 2024
- October 8, 2024
- October 1, 2024
- September 19, 2024
- September 4, 2024