About
The CTA software and service have been designed to match the Run-3 write performance requirements at WLCG Tier-0. The queuing system, coupled with a small SSD-based cache, has demonstrated its performance predictability over the first two years of the run.
This performance was achieved with FIFO scheduling, resulting in a pure temporal collocation on tapes, with only the legacy "storage class" concept to separate data into different tape pools.
As tape data is becoming "warmer" in experiment data workflows, the performance of data retrieval from tape is becoming more important. Optimising retrieval requires additional metadata to deliver staging efficiency gains, which can be exploited in the evolution of the CTA tape scheduler.
This presentation will focus on various scheduling issues observed during Run-3, and how archive metadata will enable improvements to the CTA tape site write and read efficiencies..
0:00:00 Slide 1
0:02:39 Slide 2
0:04:19 Slide 3
0:05:54 Slide 4
0:07:33 Slide 5
0:08:48 Slide 6
0:10:28 Slide 7
0:11:40 Slide 8
The talk
Credits and licence
- Institution
- CERN
- Department
- HEP Computing
- Speaker
- Julien Leduc
- Year
- 2024
- Language
- English
- Subjects
- Computer science
- Source
- CERN
- Licence
- CERN, as CERN states it. CERN's terms
- Attribution
- © CERN. CERN provides its audiovisual media free of charge for educational and informational use, credited to CERN; this lecture is shown in CERN's own player.
- Who can watch
- CERN gives no rating. Invisible Network shows it to Tweens profiles and older, judged from what the source says about it. How courses are placed