**2.1 WLCG mission**

The Worldwide LHC Computing Grid (WLCG) project [13] was launched in 2002 to provide a global computing infrastructure to store, distribute and process the data annually generated by the LHC. It integrates thousands of computers and storage systems in hundreds of data centers worldwide, see Figure 2. CERN itself provides only about 20% of the resources needed to manage the LHC data. The rest is provided by the member states' national computing centers and research network structures supported by national funding agencies. The aim of the project is the "collaborative resource sharing" between all the scientists participating in the LHC experiments, which is the basic concept of a Computing Grid as defined in [14]. The infrastructure is managed and operated by a collaboration between the experiments and the participating computer centers to make use of the resources no matter where they are located.

The collaboration is truly worldwide: it involves 35 countries on 5 continents and represents 49 funding agencies having signed the WLCG Memorandum of Understanding on Computing (WLCG MoUC) [15]. The distributed character has also a sociological aspect: even if the contribution of the countries depends on their capabilities, a member of any institution involved can access and analyze the LHC data from his/her institute.

4 Will-be-set-by-IN-TECH

When the data started to flow from the detectors, the distributed data handling machinery was performing almost flawlessly as a result of many years of a gradual development, upgrades and stress-testing prior to the LHC startup. As a result of the astounding performance of WLCG, a significant number of people are doing analysis on the Grid, all the resources are being used up to the limits and the scientific papers are produced with an unprecedented

Section 7 contains a short summary and an outlook. This chapter is meant to be a short overview of the facts concerning the Grid computing for HEP experiments, in particular for the experiments at the CERN LHC. The first one and half a year of the LHC operations have shown that WLCG has built a true, well functioning distributed infrastructure and the LHC experiments have used it to rapidly deliver Physics results. The existing WLCG infrastructure has been and will be continuously developing into the future absorbing and giving rise to new technologies, like the advances in networking, storage systems, middleware services and

As mentioned in section 1, the LHC experiments are designed to search for rare events with the signal/noise ratio as low as 10−13. This Physics requires a study of enormous number of pp and Pb-Pb collisions resulting in the production of data volumes of more than 10 PetaBytes per one data taking year. The original estimates elaborated when the LCG TDR [13] was put together were about ∼ 15 PetaBytes (PB) of new data each year which translates into ∼ 200 thousands of CPUs/processor cores and 45 PB of disk storage to keep the raw, processed and

Nowadays, 200 thousands cores does not sound like much and one can find them in large computer centers. 50 PB of a disk storage is however not that common. In any case, at the time the LHC Computing Grid was launched there was no single site within the LHC community able to provide such computing power. So, the task of processing the LHC data has been a

The Worldwide LHC Computing Grid (WLCG) project [13] was launched in 2002 to provide a global computing infrastructure to store, distribute and process the data annually generated by the LHC. It integrates thousands of computers and storage systems in hundreds of data centers worldwide, see Figure 2. CERN itself provides only about 20% of the resources needed to manage the LHC data. The rest is provided by the member states' national computing centers and research network structures supported by national funding agencies. The aim of the project is the "collaborative resource sharing" between all the scientists participating in the LHC experiments, which is the basic concept of a Computing Grid as defined in [14]. The infrastructure is managed and operated by a collaboration between the experiments and the participating computer centers to make use of the resources no matter where they are located. The collaboration is truly worldwide: it involves 35 countries on 5 continents and represents 49 funding agencies having signed the WLCG Memorandum of Understanding on Computing (WLCG MoUC) [15]. The distributed character has also a sociological aspect: even if the contribution of the countries depends on their capabilities, a member of any institution

speed within weeks after the data was recorded.

inter-operability between Grids and Clouds.

distributed computing problem right from the start.

involved can access and analyze the LHC data from his/her institute.

**2. WLCG**

simulated data.

**2.1 WLCG mission**

Fig. 2. Distribution of WLCG computing centers

Currently, the WLCG integrates over 140 computing sites, more than 250 thousands CPU cores and over 150 PB of disk storage. It is now the world's largest computing grid: the WLCG operates resources provided by other collaborating grid projects: either the two main global grids, EGI [16] and OSG [17], or by several regional or national grids.
