Cargando…
Persistent Memory based Key-Value Store for Data Acquisition Systems
<!--HTML-->Data acquisition (DAQ) systems are a key component for successful data taking in any experiment. The DAQ is a complex distributed computing system and coordinates all operations, from the data selection stage to long-term storage. With increasing throughput demand, new ways to handl...
Autor principal: | |
---|---|
Lenguaje: | eng |
Publicado: |
2019
|
Materias: | |
Acceso en línea: | http://cds.cern.ch/record/2692497 |
_version_ | 1780963948467585024 |
---|---|
author | Maciejewski, Maciej |
author_facet | Maciejewski, Maciej |
author_sort | Maciejewski, Maciej |
collection | CERN |
description | <!--HTML-->Data acquisition (DAQ) systems are a key component for successful data taking in any experiment. The DAQ is a complex distributed computing system and coordinates all operations, from the data selection stage to long-term storage.
With increasing throughput demand, new ways to handle distributed data are being explored. One of the approaches is to use specialized Key-Value store, that would be able to handle data rates bound to network interfaces only (100-400 Gbps), while providing usage flexibility with specialized APIs designed for DAQ.
We shall present how combining new storage technologies based on Intel Optane DC Persistent Memory and NVMe drives, with specialized design of Key-Value store can help achieve throughput goals, and provide operational flexibility and safety. |
id | cern-2692497 |
institution | Organización Europea para la Investigación Nuclear |
language | eng |
publishDate | 2019 |
record_format | invenio |
spelling | cern-26924972022-11-02T22:24:31Zhttp://cds.cern.ch/record/2692497engMaciejewski, MaciejPersistent Memory based Key-Value Store for Data Acquisition SystemsIXPUG 2019 Annual Conference at CERNother events or meetings<!--HTML-->Data acquisition (DAQ) systems are a key component for successful data taking in any experiment. The DAQ is a complex distributed computing system and coordinates all operations, from the data selection stage to long-term storage. With increasing throughput demand, new ways to handle distributed data are being explored. One of the approaches is to use specialized Key-Value store, that would be able to handle data rates bound to network interfaces only (100-400 Gbps), while providing usage flexibility with specialized APIs designed for DAQ. We shall present how combining new storage technologies based on Intel Optane DC Persistent Memory and NVMe drives, with specialized design of Key-Value store can help achieve throughput goals, and provide operational flexibility and safety.oai:cds.cern.ch:26924972019 |
spellingShingle | other events or meetings Maciejewski, Maciej Persistent Memory based Key-Value Store for Data Acquisition Systems |
title | Persistent Memory based Key-Value Store for Data Acquisition Systems |
title_full | Persistent Memory based Key-Value Store for Data Acquisition Systems |
title_fullStr | Persistent Memory based Key-Value Store for Data Acquisition Systems |
title_full_unstemmed | Persistent Memory based Key-Value Store for Data Acquisition Systems |
title_short | Persistent Memory based Key-Value Store for Data Acquisition Systems |
title_sort | persistent memory based key-value store for data acquisition systems |
topic | other events or meetings |
url | http://cds.cern.ch/record/2692497 |
work_keys_str_mv | AT maciejewskimaciej persistentmemorybasedkeyvaluestorefordataacquisitionsystems AT maciejewskimaciej ixpug2019annualconferenceatcern |