contributions/cms/report-cms-feb-2018.tex

 \documentclass[a4paper]{jpconf}
 \usepackage{graphicx}
 \begin{document}
-\title{The CMS Experiment at the INFN CNAF Tier1}
+\title{The CMS Experiment at the INFN CNAF Tier 1}
 
 \author{Giuseppe Bagliesi}
 
@@ -10,7 +10,7 @@
 \ead{giuseppe.bagliesi@cern.ch}
 
 \begin{abstract}
-A brief description of the CMS Computing operations during LHC RunII and their recent developments is given. The CMS utilization at Tier-1 CNAF is described
+A brief description of the CMS Computing operations during LHC RunII and their recent developments is given. The CMS utilization at Tier 1 CNAF is described
 \end{abstract}
 
 \section{Introduction}
@@ -28,7 +28,7 @@ For 2018, CMS anticipates lower average pileup conditions than those seen at the
 
 
 \section{Run II computing operations}
-During Run~II, the computing 2004 model designed for Run~I has greatly evolved. The MONARC Hierarchical division of sites in Tier0, Tier-1s and Tier-2s, is still present, but less relevant during operations. All simulation, analysis and processing workflows can now be executed at virtually any site, with a full transfer mesh allowing for point-to-point data movement, outside the rigid hierarchy.
+During Run~II, the computing 2004 model designed for Run~I has greatly evolved. The MONARC Hierarchical division of sites in Tier0, Tier 1s and Tier-2s, is still present, but less relevant during operations. All simulation, analysis and processing workflows can now be executed at virtually any site, with a full transfer mesh allowing for point-to-point data movement, outside the rigid hierarchy.
 
 Remote access to data, using WAN-aware protocols like XrootD and data federations, are used more and more instead of planned data movement, allowing for an easier exploitation of CPU resources.
 Opportunistic computing is becoming a key component, with CMS having explored access to HPC systems, Commercial Clouds, and with the capability of running its workflows on virtually any (sizeable) resource we have access to.
@@ -38,21 +38,21 @@ In 2018 CMS plans to deploy singularity \cite{singu} to all sites supporting the
 Developments to increase CMS throughput and disk usage efficiently continue. Of particular interest is the development of the “NanoAOD” data tier as a new alternative for analysis users. The NanoAOD size per event is approximately 20 times smaller than the MiniAOD data tier and relies on only simple data types rather than the hierarchical data format structure in the CMS MiniAOD (and AOD) data tier. The NanoAOD is currently in the process of a technical validation in the computing project, and physics validation by CMS analysts. It will be included in the CMS resource planning once it starts to be adopted in CMS analyses. If successfully deployed, the NanoAOD has the potential to reduce both analysis CPU and disk resources needs.
 
 \section{CMS WLCG Resources and expected increase}
-CMS Computing model has been used to request resources for 2017 RunII data taking, with total requests    (Tier-0 + Tier-1s + Tier-2s) exceeding 1870 kHS06, 149 PB on disk, and 245 PB on tape.
+CMS Computing model has been used to request resources for 2017 RunII data taking, with total requests    (Tier-0 + Tier 1s + Tier-2s) exceeding 1870 kHS06, 149 PB on disk, and 245 PB on tape.
 However the actual pledged resources have been substantially lower than the requests (1700 kHS06, 122 PB on disk, and 200 PB on tape), due to budget restrictions from the funding agencies. To reduce the impact of this issue, CMS was able to achieve and deploy several technological advancements, including reducing the needed amount of AOD(SIM) on disk and to reduce the amount of simulated raw events on tape. In addition, some computing resource providers were able to provide more than their pledged level of resources to CMS during 2017. 
 
 Thanks to the optimizations and technological improvements described before it has been possible to tune accordingly the computing model of CMS. Year-by-year increases, which would have been large in presence of the reference computing model, have been reduced substantially.
-Historically Italy contributed to CMS computing with 12\% of the Tier-1 and Tier-2 resources. The CNAF pledges for 2017  are  66 kHS06 of CPU, 5880 TB of disk, and 21 PB for tape. 
+Historically Italy contributed to CMS computing with 12\% of the Tier 1 and Tier-2 resources. The CNAF pledges for 2017  are  66 kHS06 of CPU, 5880 TB of disk, and 21 PB for tape. 
 In 2018 the CMS CNAF pledges will be 72 KHS06 of CPU, 7200 TB of disk, 24.44 PB of tape, which is still  a sizeable increase with respect to 2017.
 
-CMS usage of CNAF is very intense and it represents consistently the second Tier-1 in CMS as number of processed hours, after the US Tier-1; the same holds for total number of processed jobs, as shown in Fig.~\ref{cms-jobs}.
+CMS usage of CNAF is very intense and it represents consistently the second Tier 1 in CMS as number of processed hours, after the US Tier 1; the same holds for total number of processed jobs, as shown in Fig.~\ref{cms-jobs}.
 
 
 \begin{figure}
 \begin{center}
 \includegraphics[width=0.8\textwidth]{cms-jobs.eps}
 \end{center}
-\caption{\label{cms-jobs}Jobs processed at CMS Tier1s + Tier0 during 2017}
+\caption{\label{cms-jobs}Jobs processed at CMS Tier 1s + Tier0 during 2017}
 \end{figure}
 
 
@@ -62,21 +62,21 @@ On November 9th a major incident happened when the CNAF computer center was floo
 This caused an interruption of all CNAF services and the damage of many  disk arrays and servers, as well as of the tape library. About 40 damaged tapes (out of a total of 150) belonged to CMS. They contained unique copy of MC and RECO data.  Six tapes contained a 2nd custodial copy of RAW data.
 A special recovery procedure was adopted by CNAF team through a specialized company and no permanent data loss is expected at the time of writing.
 
-The impact of this incident for CMS, although serious, was mitigated  thanks to the intrinsic redundancy of our distributed computing model. Other Tier1 are increasing temporary their share to compensate the CPU loss, and they are trying to deploy 2018 pledges as soon as possible.
+The impact of this incident for CMS, although serious, was mitigated  thanks to the intrinsic redundancy of our distributed computing model. Other Tier 1 are increasing temporary their share to compensate the CPU loss, and they are trying to deploy 2018 pledges as soon as possible.
 A full recovery of CMS services from the  incident is expected by the end of February 2018.
 
-It is important to point out that before the incident the site readiness of CNAF in 2017 was very good and at the same level of the other CMS Tier1s, see Fig.~\ref{tier1-cms-sr}.
+It is important to point out that before the incident the site readiness of CNAF in 2017 was very good and at the same level of the other CMS Tier 1s, see Fig.~\ref{tier1-cms-sr}.
 
 \begin{figure}
 \begin{center}
 \includegraphics[width=0.8\textwidth]{tier-1-sr-2017.eps}
 \end{center}
-\caption{\label{tier1-cms-sr}Site readiness of CMS Tier1s in 2017 (excluding the last part of 2017, when CNAF was offline due to the flood incident)}
+\caption{\label{tier1-cms-sr}Site readiness of CMS Tier 1s in 2017 (excluding the last part of 2017, when CNAF was offline due to the flood incident)}
 \end{figure}
 
 
 \section{Conclusions}
-CNAF has always been an important asset for the CMS Collaboration, being the second Tier1 in terms of resource utilization, pledges and availability.
+CNAF has always been an important asset for the CMS Collaboration, being the second Tier 1 in terms of resource utilization, pledges and availability.
 The unfortunate incident of the end of 2017 has been managed professionally and efficiently by the CNAF staff, guaranteeing the fastest possible recovery with minimal data losses. The CMS collaboration looks forward to a complete recovery  and an even greater role of CNAF inside the CMS Computing in the forthcoming years.