Nowadays, Internet of Things applications face serious data and privacy protection vulnerabilities. To address some of the data protection and privacy issues, in this work we propose a new design for the self-encryption method based on a cryptographic-puzzle algorithm, that includes the generation of multiple secret keys, derived from the plaintext. As the ciphertext is constructed from several chunks of encrypted data, the absence of one of the decryption keys or one of the encrypted chunks renders recovery of the original plaintext nearly impossible. As security improvement upon to other related work proposing self-encryption, the plaintext is mixed with random values in order to use a technique known as Privacy Amplification. Privacy Amplification is achieved by applying cryptographic functions from which SHA-2 family is based on. Implementations of our design are also provided, and they are enabled for standalone and back-end execution systems. Furthermore, performance and security results and comparisons with previous related work are also provided. The security analysis confirms the use of the SHA-2 cryptographic hash family for Privacy Amplification.