Use optimised version of ABE encryption and decryption

[Manuthor]

should be done in :

• wasm32-wasi interface
• wasm32-bindgen

This optimization is actually the use of a EncryptionCache or DecryptionCache: see src/interfaces/ffi/hybrid_gpsw_aes.rs. Play with bench.rs in order to monitor perfs.

Before using this cache, it could be interesting to understand why the public key deserialization is slow (shoud we replace from_bytes with serde?)

[Manuthor]

The deserialization process is slow du to deserialization of group Gt elements. During this deserialization, cryptographic verifications are done to ensure than Gt elements are what they claim to be : the function is_gt_element, added to guaranty this property, consumes around 40% of time of this process.

Another remark concerns the WASM binaries: extract of Not So Fast: Analyzing the Performance of WebAssembly vs. Native Code:

• The instructions produced by WebAssembly have more loads and stores than native code (2.02× more loads and 2.30× more stores in Chrome; 1.92× more loads and 2.16× more stores in Fire- fox). We attribute this to reduced availability of registers, a sub-optimal register allocator, and a failure to effectively exploit a wider range of x86 addressing modes.
• The instructions produced by WebAssembly have more branches, because WebAssembly requires several dynamic safety checks.
• Since WebAssembly generates more instructions, it leads to more L1 instruction cache misses