Cloud service providers need to have elevated level of security than local system users, as users outsource their data to a remote system relying on its genuineness and also data has to be protected from intruders that exist globally. Multi-cloud systems are more prone to various kinds of attacks due to numerous internal communications of sensitive data. Multi-cloud infrastructure security can be enhanced by using stringent encryption algorithms or protocols with several integrity checks on each transaction, but that may drastically reduce the efficiency of the system. This research is one of a kind where we propose a multi-cloud framework comprising different techniques for data security as well as internal server communication security that can efficiently secure the privacy of data from eavesdroppers. Use of efficient cryptographic algorithms, modified Diffie-Hellman key exchange scheme and fragmentation technique assures overall efficiency of a system along with high security. The newest variant of SHA, i.e., SHA-3 outperformed other variants in terms of time efficiency and security. Results on multi-cloud setup demonstrate efficiency of proposed framework in terms of time taken for data storage, computations and retrieval while eliminating risks of attacks.