How do I create a distributed shared memory environment?

(I am not an expert in this field, but I will answer based on information I have access to the Internet.)

Is the technology of sharing memory in a network put to practical use?(Maybe in the research stage?)

As of 2021, it is questionable whether there is a practical "distributed shared memory (DSM)" system that is not for research purposes.The English Wikipedia page has some examples, but some projects have been completed, and I don't know if they are practical products.

Other DSM systems found:

• Omni/SCASH:Cluster-enabled Omni OpenMP on a software distributed shared memory system SCASH

I want to know the name of the technology or protocol that makes up the distributed shared memory (because it doesn't go well).

The primary purpose of Distributed Shared Memory (DSM) is to "provide application programs with a single memory space, as if they were operating on a single node without being aware of distributed processing."

DSM systems provide a variety of mechanisms on the part of dedicated hardware and middle-tier software to achieve this goal.The communication protocol between nodes is designed and selected for each specific DSM system.

Reference Links

https://www.eidos.ic.i.u-tokyo.ac.jp/research/%E5%88%86%E6%95%A3%E5%85%B1%E6%9C%89%E3%83%A1%E3%83%A2%E3%83%AA/

アプリ自体は共有メモリの形で生産的に記述しつつも， 実際には分散メモリシステム上で高速実行できるシステムを作る， という考え方が1980年代ぐらいからあって， こういう形態のシステムを「 分散共有メモリ (Distributed Shared Memory, DSM) 」と呼びます． スパコン上だと実際のところメモリは共有されていないので， その上で「あたかも共有されているかのように」見せるために システムが自動的に通信を行うわけです． DSMとは，共有メモリのインターフェイスを分散メモリシステムで 仮想化 (virtualization) する技法であると捉えられ， ハードウェアでもソフトウェアでも実装することができます． 当初は性能と生産性を両立できるシステム形態として注目されたのですが， 数十コア程度ならまだしも，そThere is a difficult situation in which researchers gradually gave up because they found it difficult to properly scale (=scale) performance with more systems.