Abstract

Emerging high-performance storage technologies are opening up the possibility of designing new distributed data acquisition (DAQ) system architectures, in which the live acquisition of data and their processing are decoupled through a storage element. An example of these technologies is 3D XPoint, which promises to fill the gap between memory and traditional storage and offers unprecedented high throughput for nonvolatile data. In this article, we characterize the performance of persistent memory devices that use the 3D XPoint technology, in the context of the DAQ system for one large Particle Physics experiment, DUNE. This experiment must be capable of storing, upon a specific signal, incoming data for up to 100 s, with a throughput of 1.5 TB/s, for an aggregate size of 150 TB. The modular nature of the apparatus allows splitting the problem into 150 identical units operating in parallel, each at 10 GB/s. The target is to be able to dedicate a single CPU to each of those units for DAQ and storage.

Highlights

  • O VER the last years, large-scale computing systems are moving toward a direction where compute and storage capabilities are decoupled [2], [3]

  • In the context of data acquisition (DAQ) systems for Physics experiments, the decoupling of acquisition and processing is interesting in those cases in which the acquisition rate may vary widely in time depending on the physical processes being measured: an intermediate storage element allows to dimension the data processing part of the system for an average load without needing to sustain temporary peaks

  • Intel Optane DC Persistent Memory modules have been tested in detail with both synthetic benchmarks and with a custom-made high-level application

Read more

Summary

INTRODUCTION

O VER the last years, large-scale computing systems are moving toward a direction where compute and storage capabilities are decoupled [2], [3]. Emerging high-performance storage technologies are being used in the design of new distributed DAQ system architectures where data production and data processing are decoupled by a large storage buffer. An example of these technologies is Manuscript received November 11, 2020; revised March 6, 2021 and April 11, 2021; accepted May 23, 2021. One possible way to achieve the target objective for the system is to use fast storage media An example of such devices is the Intel Optane Data Center Persistent Memory Modules (DCPMMs) that leverage the 3D XPoint technology. The suitability of DCPMMs for the DUNE storage buffer is explored, emulating the workflow on a prototype setup at CERN [10]

INTEL OPTANE
Operation Modes
RELATED WORK
System Description
Testing Strategy
BENCHMARKS OF THE DCPMMS AND DISCUSSION
APPLICATION FOR THE DUNE DATA ACQUISITION SYSTEM AND DISCUSSION
CONCLUSION
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call