Multilevel Encryption for Multiple-Cloud Storage (MEMC) with Bot Detection and Elimination (BDE) in Recommendation Systems

Abstract

Introduction: In general cloud acts as an intermediate in Publish and Subscribe (pub/sub) systems to broadcast the publisher’s data to the subscribers. But a direct connection between the subscriber and the publisher is not advisable, only a loosely coupled manner allowed along with the cloud. There are a lot of chances in data getting exposed to the attackers from the client side, publisher side or even from the cloud side. Attackers, specifically botnet controllers, use stealthy commanding systems to set up large-scale controls. Encryption before upload is the best way of protecting sensitive publications and subscriptions.

Objectives: To suggest a security system using Multi-level Encryption mechanism for multiple cloud storage(MEMC)  with BotDetection and Elimination(BDE) strategy for recommendation systems.

Methods:. A privacy-preserving publish and subscribe system that protects both the publications and the subscription's personal details and identities is presented.  The cloud storage that compromises the pub and subsystem is broken using a high multi-level encryption approach. A novel paradigm emerges from combining multi-level encryption with multiple cloud storage (MEMC).A 'Searchable Encryption (SE)' technology is used to ensure encrypted publishing keyword matching, in addition to subscriber interest. The system's efficiency is based on employing various cloud storage for matching and directing trustworthy publications to trustworthy interested subscribers. The optimum solution is to divide the match procedure into multiple phases, coupled with encrypted subscriptions and publication tags. • The partitioned data will be handled by various cloud servers to prevent sensitive data leaks. Even if a single cloud server is attacked and collaborates with a subscriber or publisher to leak data, data privacy is unaffected; subscriptions and publications remain protected. The system uses the 'BOT detection and Elimination (BDE)' methodology to detect and stop BOT control.

Results: MEMC clearly outperforms PEKS, SE, and AES in the Confidentiality criteria, with a score above 3.5. This suggests that MEMC is better suited to maintaining data confidentiality. For the Privacy Maintenance metric MEMC significantly excels, with a value close to 4, showing that the method is effective in the maintenance of user privacy with multilevel encryption. The other methods such as PEKS and AES show relatively lower privacy scores, reflecting their inability to present a solution to the enhanced security concerns

Conclusions:The proposed mechanism is used to secure the publications and subscriptions by providing a multilevel encryption technique before it is disseminated to the cloud. To tackle the untrusted cloud the system provides an Epub/Esub encryption system to maintain confidentiality. Further, the BOT type attacks are common in clouds to resist and overcome this system and  introduces a BOT detection and elimination technique. Finally, the experimental results ensure the efficiency, performance and feasibility of the system.