While blockchain project grows, resources will be in scarcity: network, memory, storage and CPU. Sharding is a vision proposed by many projects to solve scaling issues by splitting jobs to different nodes. It is commonly used in database technology that you organize many machines to store big data by split data to different layers.
Recently Litecoin founder Charlie tweeted: by definition, a decentralized cryptocurrency must be susceptible to 51% attacks whether by hashrate, stake, and/or other permissionlessly-acquirable resources. If a crypto can’t be 51% attacked, it is permissioned and centralized.
In this spirit, I believe that sharding falls into “trust and centralized” space. Sharding requires to trust other nodes for completing the job. As soon as one has to trust another entity rather than “math proof”, it is no longer trustless blockchain. Bitcoin trust-less nature means only trusting in math. Once trust “other” structure is setup, it will soon be centralized.
So far “permission-less” and “trust-less” original principles are preventing blockchain to scale up for good blessing which is fight against centralization. It has always be a nightmare for TAU on the speed up TPS while maintaining these two. EOS is trust-less protocol super fast in TPS without “shards” but it has permission requirements for 21 super nodes which is easily under government control. ETH 2.0 tries to stay in trust-less and accomplish “sharding”. It has taken too long that I am concerned that it might never get done.
For TAU strategy, I was researching on scaling by using “shard” concept. Thanks to Charlie, I think it might not be a right strategy to trade “trust” to get “shards” for increasing the speed. We have many high speed database, but very few trustless and permission-less database. We have to find other way or simply just wait on our civilization to increase the network speed for higher transaction volume, such as mobile video has to wait until 4G.
We all love speed, but we love “trustless and permissionless” more.