Web3 compute has a trust issue, but the solution is obvious

Web3 is powering a digital revolution that will bring untold benefits to organizations. Decentralization promises to tear down the monolithic structures that support the internet as it exists now, with major implications for finance, social media, and even the computing infrastructure that supports the digital economy. 

The prospect of decentralized compute has many eager with anticipation, because it can bring incredible cost benefits by utilizing idle computing resources while preventing censorship. More than that, it can provide enhanced scalability for artificial intelligence workloads, and it supports ideals around privacy and sovereignty, giving users full control over their data. 

But there’s one outstanding challenge we must overcome before we can fulfill this decentralized dream — namely, the need to establish trust in decentralized compute. The question is, how can this be done without the assurances provided by cloud computing giants such as Amazon Web Services or Google Cloud? 

Those legacy cloud computing giants dominate the compute industry, even while charging exorbitant prices for their services and having questionable track records in terms of data privacy, simply because of the trust they command. By offering service level agreements within a clear, hierarchical structure, users are assured that they’re getting the reliable, scalable compute they need to power their applications. If you pay for premium uptime, guaranteed performance, and dedicated support, you know that if they don’t deliver it, you’ll have legal recourse. 

Today’s cloud giants operate in a framework that enables contracts to be enforced. Users know that downtime is an anomaly, and on the rare occasions when it happens, they’ll be compensated for the problems caused. And if that compensation isn’t forthcoming, they have clear avenues to seek recompense. This is why centralization is so powerful. Despite its limitations, it provides strong assurances and accountability, which means protection for users.

Trust is critical

As the crypto industry pushes the shift to web3 infrastructures and decentralized compute, this centralized model of trust doesn’t apply. After all, web3 seeks to kill these intermediates and single points of failure, and redistribute power equally among its users, and that means there’s no obvious recourse in the event that problems occur. Although it’s an immensely exciting shift, it leads to questions about how trust can be enforced. If web3 cannot establish trust, it’s unlikely to be able to displace centralized providers in an industry as critical as cloud-based compute.  

Instead of one massive data center operated by a rich and powerful corporation, decentralized networks have thousands, if not millions, of individual nodes, each contributing a little bit of power to the network. By combining these resources, it’s possible to make immense computing resources available to those who need them at lower costs, but those users require assurances, too. 

For instance, a cash-strapped AI startup seeking a cluster of powerful GPUs is likely to find the idea of an affordable decentralized compute network appealing, but how can it know for sure that the resources it’s paying for are reliable? How can it verify their computations? In a network where anyone can contribute resources, how can it identify which nodes are reliable and trustworthy, and which ones might be slow and potentially even malicious? 

The web2 model, based on enforceable SLAs and brand recognition, simply doesn’t apply to decentralized networks. In fact, the very idea is anathema to web3, because if you had a single entity that’s able to enforce whatever guarantees are made, that means having to accept the lack of privacy and the potential for censorship it promises to eradicate. 

The issue of trust is a critical one that must be solved; otherwise, decentralized compute’s growth will be handicapped by a lack of confidence. An application that has millions of users globally needs to know it can rely on its underlying servers, and if web3 can’t offer any assurances, it will have little choice but to rely on centralized infrastructure providers due to the strong guarantees they provide, even if their model undermines its own, decentralized principles. 

Building community trust with incentives

Fortunately, web3 offers an elegant solution that aligns with its core ethos. The answer is to engineer trust through a system of decentralized audits by incentivized, community-run validator nodes. 

So instead of having compute nodes that are vouched for by an organization like AWS, which can be sued if it breaks its promises, web3 must instead rely on the collective intelligence and vigilance of hundreds of network participants, rewarding them for their honesty and penalizing them for not telling the truth. 

The individual validators, of which there could be thousands, can be incentivized to act honestly via reward-based staking mechanisms. This will encourage them to accurately assess and verify the performance and reliability of each node. Collectively, these validators will monitor the entire network of compute providers, auditing them on a continuous basis. Their job will be to verify the correctness of their computations, measure their performance, latency, and uptime, and identify any nodes engaged in malicious behavior. Users will then be able to look at the overall consensus, and in this way, the validators generate trust in the network. 

To encourage positive behavior, a “carrot-and-stick” approach is used. Should any compute node fail to meet the expected level of performance or attempt any funny business, it would be quickly identified by validators and penalized, taking away any incentives it may have. Meanwhile, the best-performing nodes will be rewarded, enhancing their reputations and attracting more demand for the services they provide. Moreover, the validators themselves will be penalized or rewarded, based on their honesty. 

Anyone who knows anything about crypto will immediately recognize the validity of this model, for it’s already used in countless proof-of-stake blockchains, where validator nodes work together to verify transactions. With decentralized compute, these validators will instead verify computations, creating a transparent and tamper-proof system of trust that’s every bit as reliable as the SLAs offered by AWS. 

A superior trust foundation

Decentralized audits by validator nodes align perfectly with the web3 model. It’s a permissionless model, and just as everyone can provide compute to the network, anyone can become a validator, meaning it’s fair to every participant. Moreover, the audits are completely transparent, with their processes and results published on the blockchain for anyone to verify. 

The design of such a system means it’s in the best interests of every validator to act honestly, as they’re incentivized to maintain a reputation for honesty, lest they lose their rewards and forfeit their stake. 

Building such a framework is challenging, no doubt, with the need for robust verification algorithms, easy-to-understand trust profiles, and simple requirements for users to become validators and join in the process. But once these frameworks are up and running, decentralized compute networks will be able to offer a superior foundation of trust and move beyond the limitations of today’s centralized cloud providers.