Abstract

A many-to-one communication pattern is present both in Data Acquisition (DAQ) and datacenter networks. The problem arising from this pattern is widely known in the literature as incast and can be observed as TCP throughput collapse. It is a result of overloading the switch buffers, when a specific node in a network requests data from multiple sources. This paper provides two contributions. First, we confirm that there are strong analogies between the TCP behavior in DAQ and datacenter networks. Second, we evaluate different proposals from the datacenter environment for application in DAQ to improve performance and reduce buffer requirements.

Highlights

  • ATLAS [1] is a general-purpose particle detector designed to study particle collisions at the Large Hadron Collider (LHC) at CERN

  • The ATLAS Data Acquisition (DAQ)/HLT system is based on 10 Gbps Ethernet network with TCP as transport layer protocol and is capable of filtering 2 Tbps of event data, which is required by the experiment

  • In this paper we showed that there are analogues between the use of TCP in DAQ networks and the incast problem in datacenters

Read more

Summary

INTRODUCTION

ATLAS [1] is a general-purpose particle detector designed to study particle collisions at the Large Hadron Collider (LHC) at CERN. ATLAS, like other existing and emerging largescale experiments, produces increasingly high volumes of data Their source can be a wide variety of instruments, including sensors, detectors, antennas or telescopes. The ATLAS DAQ network is based on TCP/IP and Ethernet technologies [3] This poses, several challenges due to the bursty nature and many-to-one communication pattern of the data flow in DAQ. It resembles one of the patterns identified to be very common in typical datacenter networks: incast [4]. In the first part of the article we describe a typical DAQ network and reveal its relation to the TCP incast in datacenters. In the second half of the article we validate proposals for solving datacenterincast in the real DAQ system of the ATLAS experiment

Data Acquisition Networks
Relation to TCP incast in datacenters
Related work
CURRENT SYSTEM
Network configuration
Traffic shaping
APPROACHES TO AVOID TCP-INCAST
Increasing the buffer sizes
Fine-grained TCP retransmissions
Static TCP configuration
DIATCP
Findings
CONCLUSION

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

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.