Storage · Memory
08-25[memory]Mobile Memory Takes the Server Socket: LPDDR6 Silicon, SOCAMM2 and DRAM That Computes19m08-24[memory]The Boundary Learned to Move: 52 Industry Memory and Storage Papers in 202628m08-03[memory]NAND Moves Up a Tier: Reading the First High Bandwidth Flash Specification20m07-13[memory]Memory Capacity Is Not Memory Bandwidth: RamRyder Gives Cloud VMs Their Own Channels17m07-13[memory]A Cache Hit That Arrives Late Is Still a Miss: Strata Schedules Long-Context Memory19m06-27[memory]CXL Leaves the Lab: How Meta Turned Retired DDR4 into a Production Memory Tier21m05-04[memory]Before the GPU Reads a Byte: FalconFS Moves the Directory Walk Off Training Clients17m05-04[memory]No Giant Switch: How Octopus Builds a 96-Server CXL Memory Pod13m03-24[memory]A Real CXL Memory Box Under SAP HANA: What Can Move, and What Must Stay Local13m02-24[memory]Five Regions Changed the Economics of Apple's Object Store16m02-24[memory]KV Reuse Breaks When the Text Moves, Even If the Text Does Not17m02-24[memory]The SSD Stayed Local, but Its Control Path Moved Three Times17m02-24[memory]The Fastest Model Loader Changed the Page Cache, Not the Framework16m02-24[memory]Compression Improves When the Filesystem Sorts Before It Packs14m02-24[memory]A Thousand Cartridges, Four Drives: Why Tape Cloud Storage Must Wait on Purpose19m
# 2025
11-25[memory]KVCache Without the Network: Alibaba Runs LLM Serving on a Real CXL 2.0 Switch19m11-03[memory]The Hierarchy Became the Product: 34 Industry Memory and Storage Papers from 202529m07-07[memory]Read Layout Is Not a Recovery Group: Okapi Separates Two Storage Decisions18m05-01[memory]Eighty Percent Far, Eighty-Five Percent Fast: A Database Placement Rule for CXL12m04-28[memory]RDMA Remote Memory Without a Memory-Node CPU15m02-16[memory]HBM Saves Clock Power by Rebuilding Four Phases at the Data Groups24m# 2024
08-06[memory]Memory Stopped Being a Component: 35 Industry Papers from 202429m