Experience running a distributed Tier-2 in Spain for the ATLAS experiment

The main role of the Tier-2s is to provide computing resources for production of physics simulated events and distributed data analysis. The Spanish ATLAS Tier-2 is geographically distributed among three HEP institutes: IFAE (Barcelona), IFIC (Valencia) and UAM (Madrid). Currently it has a computing power of 430 kSI2K CPU, a disk storage capacity of 87 TB and a network bandwidth, connecting the three sites and the nearest Tier-1 (PIC), of 1 Gb/s. These resources will be increased according to the ATLAS Computing Model with time in parallel to those of all ATLAS Tier-2s. Since 2002, it has been participating into the different Data Challenge exercises. Currently, it is achieving around 1.5% of the whole ATLAS collaboration production in the framework of the Computing System Commissioning exercise. A distributed data management is also arising as an important issue in the daily activities of the Tier-2. The distribution in three sites has shown to be useful due to an increasing service redundancy, a faster solution of problems, the share of computing expertise and know-how. Experience gained running the distributed Tier-2 in order to be ready at the LHC start-up will be presented.