Most of us are well aware that bringing a project to completion is an exercise in priority and compromise. In project management this is encapsulated with the Iron Triangle model, which can be summed up as: cost, speed, quality — choose two.
Or, if you want a project done fast and cheap, the quality will suffer; if you want it done fast and at high quality, it will cost you. Moving from philosophy to application, the blockchain industry too is coming to terms with its own priority model.
Vitalik Buterin, the creator of Ethereum and one of the most influential people in the cryptocurrency world, has posited a Trilemma model which lists decentralization, scalability, and security as the major vectors. As with the Iron Triangle, you can have two, but not all three. For example, Bitcoin and Ethereum are both fully decentralized and secure — because every node stores a copy of all the data, and no one entity controls the network — but the fact that each node stores all of the data mean both cryptocurrencies are limited in scale.
The Tatau Trilemma
In developing a reliable and usable enterprise platform, Tatau too has prioritized the Trilemma factors most critical to us and our customers.
First and foremost is security — we won’t be making any compromises when it comes to the security of our customers’ data. The Tatau platform will be handling advanced and sensitive data, so it’s critical to protect it against breaches and maintain privacy.
Scalability is also a non-negotiable because limiting the scale would limit the scope of our customers. Tatau provides fast and secure processing for intensive AI computing and predictive analytics — to deliver on that promise our platform needs to reach a global scale.
So what about decentralization? Although it is the buzzword in the blockchain industry, it doesn’t always make sense for a platform to be fully decentralized — at least not yet. To deliver the non-negotiable requirements of our customers, today, the parts of our platform that handle job management and verification need to be handled by Tatau. That being the case, we have designed the Tatau platform to be distributed, maximizing utility and scalability, but not entirely decentralizing the critical parts.
Secure Job Management
When a customer comes to Tatau, we receive the data they want to be processed and break it into chunks for distribution to our distributed network of suppliers. Once the work is done, we consolidate and deliver the result to the customer. Job management is a relatively ‘light’ process, using a minimal amount of resources, but it is also extremely important.
Decentralizing this process would bring an increased risk of a security or privacy breach. We don’t believe it’s worth the risk, for us or our customers, so we have elected to manage the data in a secure environment and protect it with advanced encryption before we distribute it. Therefore, it makes more sense for job management to be performed by a trusted entity — hence the Tatau platform initially performs all job management.
In saying this, things change and we have an exciting roadmap ahead of us, including future plans to create an SDK so that customers can self-manage their own jobs if they choose to. This logical future would ensure the continued health and scalability of the Tatau ecosystem. It’s not so much ‘decentralization’, but rather a ‘self-custody’ strategy — further empowering the customer.
One of the things that sets Tatau’s platform apart is the ability to validate when complex computations have been completed and done correctly. We use machine learning to verify that results are consistent and valid across all computation nodes.
In an open network, there’s always a risk of fraud and collusion by ‘bad actors’, especially one where money is changing hands. You could imagine that a supplier node could try to get paid for a job that it did not complete correctly. Therefore, initially, Tatau will handle verification using a distributed but server-based approach, so that we can ensure the process is secure and robust. By verifying the work ourselves, we remove the risk of paying people for work they didn’t do or failed to do correctly.
At this stage, fully-decentralizing the verification process doesn’t offer our customers a benefit — on the contrary, it would add a level of uncertainty. As a result, Tatau maintains quality control over this part of the process.
We see a possible future where this can become a decentralized process too but when working on the cutting edge of technology, we need to know the responsible limits and stay well within them when we’re dealing with enterprise customers and sensitive data.
Solving the Trilemma
Blockchain is a young technology. Every product we currently see in the market is still in development and is largely untested. But because it’s such a young technology, there is plenty of room for improvement. The limits of this technology can be pushed so perhaps eventually, instead of choosing between decentralization, scalability, and security, we can have all three. It’s one of many problems we’re working to solve here at Tatau — and look forward to engaging with our community about it in the near future.