The challenges and solutions faced by FHE are different from ZK and MPC, and FHE is still in the early stages.
The main challenges include:
Slow performance: Currently, private smart contracts using fh-EVM have only 5 TPS. Additionally, TFHE's current performance is about 1000 times slower compared to pure data.
Not developer-friendly enough: There is still a lack of standardized algorithms and overall supported FHE tools.
High computational overhead (cost): The management of noise and the complexity of computations can lead to node centralization.
Unsafe on-chain FHE risks: For any secure threshold decryption system, the decryption keys are distributed among nodes. However, due to the high overhead of FHE, this may result in a reduced number of validators, thus increasing the likelihood of collusion.
Solutions:
Programmable bootstrapping: It allows computations to be applied during bootstrapping, thus improving efficiency while catering to specific applications.
Hardware acceleration: Develop ASICs, GPUs, and FPGAs alongside the OpenFHE library to accelerate FHE performance.
Better threshold decryption systems: In short, to make on-chain FHE more secure, we need a system (which can be MPC) to ensure: low latency, reduced decentralization entry barriers for nodes, and fault tolerance.
