Paper Title: Improving Wear Leveling on Android Smartphones

Paper Abstract: Android smartphones are a phenomenon, with over 130 million Android users in the world, and flash is the NVM storage media in Android smartphones. While flash is great for fast random I/O workloads, wear leveling has a large impact on the lifetime of the underlying flash media. This talk will address how Samsung developed a patented technology for converting small random I/Os into large sequential I/Os on Android devices, its impact on the wear of the flash, and its impact on the load and store times of applications such as Camera, App store, etc. We will focus on techniques such as compression and deduplication to reduce the storage footprint on the flash, and also changes to the SQLite database to avoid frequent flushing of data. As such, these changes allowed many of the I/O requests to be satisfied from the buffer cache on these Android devices.

Paper Author: Tejas Chopra, Sr. Software Engineer, Netflix

Author Bio: Tejas Chopra is a Sr. Software Engineer at Netflix, where he works on the Storage infrastructure for Netflix Studios. He is a TedX and an international keynote speaker with talks on Blockchain, Storage, Web3.0, Cloud computing, and Engineering culture. He has over a decade of experience working on distributed systems, and has a Masters in Electrical and Computer Engineering from Carnegie Mellon University.