Quantitative Analysis of Data Caching for the HL-LHC Data-lake

Given the development of the High Luminosity Large Hadron Collider ( HL-LHC ), the Worldwide LHC Computing Grid ( WLCG ) will face unprecedented computing challenges. The amount and complexity of the data generated from the different experiments at CERN will increase, and so the WLCG has developed a strategy to handle this through caching and the creation of data-lakes. This paper focuses on a quantitative analysis of this caching by exploring the relationship between cache size and hit-rate. This, through a simulation based on the interactions of a Tier 2 site and the Data Center. The results that show that, at least for these two sites, increasing the cache’s capacity above 0.45 petabytes will not increase above a 41% hit-rate. These results are based on the Data Center’s interactions registered in log files from May 31 to June 25, 2018.