一句话：指出基于虚拟内存的分离式内存软件运行时存在高缺页开销与页粒度脏数据放大两大局限。
分离式内存可缓解数据中心资源分配效率低下的问题。为使其实用，已有多种软件运行时借助虚拟内存子系统，以本地内存抽象向应用透明地提供分离式内存。然而，论文指出这种设计存在多重局限：一方面需借助缺页异常识别应获取并缓存在本地的数据，带来高开销；另一方面，缓存数据变更跟踪采用页粒度（4KB 或更大），导致严重的脏数据放大。该文因此主张重新思考面向分离式内存的软件运行时设计，避免依赖虚拟内存机制。
Disaggregated memory can address resource provisioning inefficiencies in current datacenters. Multiple software runtimes for disaggregated memory have been proposed in an attempt to make disaggregated memory practical. These systems rely on the virtual memory subsystem to transparently offer disaggregated memory to applications using a local memory abstraction. Unfortunately, using virtual memory for disaggregation has multiple limitations, including high overhead that comes from the use of page faults to identify what data to fetch and cache locally, and high dirty data amplification that comes from the use of page-granularity for tracking changes to the cached data (4KB or higher).
@inproceedings{paperbot6120,
title = {Rethinking software runtimes for disaggregated memory},
author = {Irina Calciu and M. Talha Imran and Ivan Puddu and Sanidhya Kashyap and Hasan Al Maruf and Onur Mutlu and Aasheesh Kolli},
booktitle = {ASPLOS},
year = {2021},
doi = {10.1145/3445814.3446713}
}