π€ AI Summary
This work addresses the inefficiency in tiered memory systems where fragmentation of hot data causes cold data to remain stranded in expensive DRAM, leading to significant resource waste. To mitigate this, the authors propose an address space engineering approach that dynamically reorganizes virtual memory by co-designing an object-aware user-space frontend with a page-aware OS backend, thereby coalescing hot and cold objects into separate physical pages. The system employs lightweight pointer instrumentation to track accessηεΊ¦ (access frequency) and leverages a lock-free protocol to enable fine-grained, concurrency-safe runtime object migration, all while remaining compatible with existing memory management mechanisms such as kswapd and Transparent Memory Offloading (TMO). Experimental results across diverse data structures and production workloads demonstrate 2β4Γ improvements in page utilization, up to 70% reduction in DRAM footprint, and only 2β5% performance overhead.
π Abstract
Hardware and OS mechanisms for memory tiering are widely deployed, yet datacenters still overprovision DRAM. The root cause is hotness fragmentation: allocators place objects by size rather than access pattern, so hot and cold objects become interleaved within the same pages. A single hot object marks its page as active, trapping surrounding cold data in expensive DRAM. Our analysis of Google production workloads shows that up to 97% of the bytes in active pages are cold and unreclaimable. We propose address-space engineering: dynamically reorganizing virtual memory so that hot objects cluster into uniformly hot pages and cold objects into uniformly cold pages. We present OBASE, a compiler-runtime system for unmanaged languages that serves as an object-aware frontend for page-aware OS backends. OBASE tracks accesses via lightweight pointer instrumentation and migrates objects at runtime using a lock-free protocol that is safe under concurrency. By reorganizing the address space, OBASE enables unmodified backends (kswapd, TMO, TPP, Memtis) to tier memory effectively. Across ten concurrent data structures, six backends, and production traces from Meta and Twitter, OBASE improves page utilization by 2-4x and reduces memory footprint by up to 70%, with only 2-5% overhead.