Big data storage centers became important in the recent era due to the advances in today's world cloud systems and the increased amount of the data generated by users. Mass-inexpensive storage devices are still used in order to provide storage for a lower cost. It is a fact that data retrieval from commonly used massinexpensive storage devices (i.e. HDDs and Tapes) consumes much power due to their cost-inefficient hardware architecture (i.e. arm moves over different tracks on magnetic disk) especially with randomly stored data blocks. Retrieving sequentially stored data blocks tends to consume less power as they require less arm moves. Hence, the need for energy efficient solutions is important. In this research, we present an energy efficient solution for big data storage system using sequential caching approach. It is practically proven that users' I/O requests for data tend to show repeating patterns. Data caching that keeps the most frequently accessed data blocks showed efficient performance improvement for many solutions over the last decades. Our Solution (EEHSC) boosts data caching and arm moves' space locality by buffering most frequently used data blocks in the middle of the storage medium. This in result reduces power consumption by increasing the percentage of data blocks that are read sequentially over those that are read randomly. We run performance evaluation for our EEHSC algorithm using a simulator created by ourselves based on real world I/O traces. Results show an EEHSC can efficiently reduce power consumption without the need for allocating big size cache due to the high data hit-rate that can be achieved in the cache. Results showed that power saving ratio can reach 67.60% when using LASR machine01 trace and 36.74% when using LASR machine06 trace. This difference goes to the variation of the repeated pattern for the most frequently used data blocks.