Next: Acknowledgements Up: Mercy: a fast large Previous: Design of Mercy

Conclusions

Large blocks are useful in any application where random access to data is desirable, of which sector encryption is a prime example. Mercy is intended to demonstrate the possibility of building an efficient block cipher for large blocks. Mercy's design was inspired by two beliefs:

• Large block sizes can lend useful advantages in security and speed
• Avalanche across large blocks need not be perfect before a cryptanalyst with limited computing resources cannot distinguish it from perfect, as explained in Section 5.6.

However, the primary motivation for the design of Mercy is not that the cipher be directly used for this application; it is to inspire further work where it is badly needed. Disk encryption suffers none of the barriers to adoption from interoperability suffered by (for example) email encryption. But it is very rarely used, and the main barrier to adoption among security conscious users is the high performance penalties it currently exacts. I hope that Mercy demonstrates that fast, secure large block ciphers are possible, and inspires better cryptographers to propose better designs.