**2.1.2 New computing-experimental tools2**

For new computing-experimental tools, we have worked on a Belle II data handling system. The Belle II experiment will begin at KEK in 2015. Belle II computing needs to include raw data reconstruction, data reduction, event simulation, and user analysis. The Belle II experiment will have a data sample about 50 times greater than that collected by the Belle experiment.

Therefore, we have very large disk space requirements and potentially unworkably long analysis times. Therefore, we suggested a meta-system at the event-level to meet both requirements. If we have good information at the meta-system level, we can reduce the CPU time required for analysis and save disk space.

The collider will cause the computing requirement for data analysis and Monte Carlo production to grow larger than available CPU resources. In order to meet these challenges, the Belle II experiment will use shared computing resources as the Large Hadron Collider (LHC) experiment has done. The Belle II experiment has adopted the distributed computing model with several computing processing systems such as grid farms (Kuhr, 2010).

In the Belle experiment (Abashian et al., 2002), we use a metadata scheme that employs a simple "index" file. This is a mechanism to locate events within a file based on predetermined analysis criteria. The index file is simply the location of interesting events within a larger data file. All these data files are stored on a large central server located at the KEK laboratory. However, for the Belle II experiment, this will not be sufficient as we will distribute the data to grid sites located around the world. Therefore, we need a new metadata service in order to construct the Belle II data handling system (Kim, et al. 2011; Ahn, et al., 2010).

Fig. 6. Data handling scenario at the Belle II experiment.

 2 This section is based on the paper titled "The embedment of a metadata system at grid farms at the Belle II experiment" by S. Ahn et al. in Journal of the Korean Physical Society, Vol. 59, No. 4, pp. 2695-2701, (2011).

The e-Science Paradigm for Particle Physics 85

Feynman diagram of Left-Right models for the analysis of *CP* violating dimuon charge

Fig. 8. The Feynman diagram of Left-Right models for the analysis of *CP* violating dimuon

In 2008, CDF showed a possible anomaly in the forward-backward asymmetry of the top quark, where AFB = 0.19± 0.07(stat.) ± 0.02(syst.) (Aaltonen et al., 2008). We have performed model independent analysis. Considering the s-, t-, and u- channel exchanges of spin-0 and spin-1 particles whose color quantum number is a singlet, octet, triplet or sextet, we study the region consistent with the CDF data at a one sigma level. We show the necessary conditions for the underlying new physics in a compact and effective way when those new particles are too heavy to be produced at the Tevatron. However, the results still affect the

For theory-computing, we study flavor physics based on lattice gauge theory, which enables large-scale numerical simulations on a supercomputer. The theory of strong interactions in the Standard Model is Quantum Chromo Dynamics (QCD). In phenomena related to the Cabibbo-Kobayashi-Maskawa (CKM) matrix, the theoretical values of the interaction amplitudes also have factors that cannot be obtained in a perturbative way since the strong coupling constant becomes strong at a low energy scale as QCD, as a non-abelian gauge theory, predicts. The only way that one can calculate the non-perturbative quantities with a controlled error is the lattice method, in which we put strongly interacting particles, quarks and gluons, on a lattice and calculate quantities directly from first principles. Fig. 9 shows

We use the staggered fermions, which are one of the more popular lattice fermion schemes for full QCD lattice simulations. The staggered fermion scheme has the advantage that its

asymmetry due to *B* mixing.

charge asymmetry due to B mixing.

forward-backward asymmetry of top quark.

**2.3 For theory-computing** 

the baryon based on lattice QCD.

**2.2.2 The forward-backward asymmetry of top quark pairs** 

Fig. 6 shows the Belle II data handling system scheme. First, a user makes a metadata query to the server. Second, the server gives back a list of files and events. Third, the server may give a list of grid sites. Fourth, the user requests job execution at grid sites. Fifth, a logical file catalog (LFC) maps a logical file name (LFN) into a set of physical file names (PFN). Finally, the computing farms at the grid site read the requested physical file (Ahn, et al., 2011).
