Member Of Technical Staff
San Francisco, California, Us
1. Work on Riverbed's enterprise storage solution called SteelFusion.2. Develop and maintain SteelFusion's core IP, i.e. the prediction algorithms that enable read IO performance across WAN.- SteelFusion uses block-level read ahead algorithms that underpin its patented BlockStream technology to consolidate branch data at the data center while projecting a working set to the branch.- Developed read ahead support for VMFS 6 and SteelFusion's NFS export configurations.- Developed and tested core in-memory structures for NFS configuration support.- Got introduced to VMFS and ext3 internals.- Analyzed memory overhead for various components of the module.3. Developed a new file system agnostic read performance engine based on FP-Growth algorithm, a Data mining algorithm proposed by:J. Han, H. Pei, and Y. Yin. Mining Frequent Patterns without Candidate Generation. In: Proc. Conf. on the Management of Data (SIGMOD’00, Dallas, TX). ACM Press, New York, NY, USA 2000- This system is a proposal for next-gen read ahead algorithms that promise re-usability, ease of development and maintenance.- Undertook C++ implementation of FP-Growth along with unit tests. Integrated the algorithm with SteelFusion's data path.- The algorithm can be re-purposed as an intelligent caching software or applied to completely different problems requiring pattern recognition.4. Worked on initiatives like utilizing cloud object storage as a backend storage for SteelFusion, utilizing REST interfaces.5. Developed SteelFusion's Virtual Machine Aware Backup solution.- VMAB consolidates backup and recovery process for all the branch offices into a single datacenter.- It simplifies backup process and generates CAPEX and OPEX savings.6. Periodically handled crashes/bugs reported by the customers.7. Improved SteelFusion's unit test infrastructure.- Introduced containers in the test infrastructure.- Introduced Google Test Framework.8. Tech: C++, Python, Unix like environment.