Researches

HPDS Corp. has been always relying on up-to-date system research to advance its products. The research and innovation is the main focus of HPDS Research. Since its foundation, HPDS Research has focused on innovative design and architectures for storage management products and solutions including SANs, NAS, data backup, and video surveillance. With the aim of innovative architectures and data management techniques, HPDS has offered various SAN products including SAB series to meet customers reliability and performance requirements with affordable cost. With high quality research activities in HPDS Research, the company has been granted various international patents in data management techniques and has also published very high quality papers in top venues.

 In HPDS Research, we make the core technologies for future generation of HPDS storage systems. To make this happen, we do research in all layers (with more focus on system software layers). Example topics of interest are as follows:

Data Reduction Technologies
1.    Data Reduction Estimation
We have developed a simple but highly practical tool to estimate data reduction (deduplication and compression) potential on a block device. It is applicable on a block device (single disk, array of disks, a LUN, etc.) and has a couple of configurable parameters (including system DRAM usage). We developed it as part of a bigger project in the domain of data reduction technologies.
https://github.com/HPDSResearch/data_reduction_savings_estimator
2.    Scalable Data Reduction
Designing a scalable data reduction operation is very important for a high performance all flash storage system. In one of our joint research projects (with academia) published in ACM SIGMETRICS’22, we designed a scalable data reduction architecture that provides up to 12x speedup over the baseline all-flash system, and improves the performance per cost by up to 57x. The source code is downloadable in the following link:
<Link> (Coming soon)

Scalable Caching:
We have open-sourced the scripts of the core part of our proposed framework published in ASPLOS’23 paper. These scripts enable analyzing the performance behavior of a desired I/O cache (esp. OpenCAS, EnhanceIO, and DM-Cache) running on top of RAM Disk (as a very fast cache device), and an SSD-array (as fast backend device).
https://github.com/HPDSResearch/io_cache_performance_analysis_framework