122018 atom c
Previous: Faculty of Nuclear
Our target is to create a distributed scaleable computer farm with the basis in the Institute of Physics AS CR. The Institute has already infrastructure and experience with data processing. The computer connection between the institutes is on the level of 64 kbps and should be enhanced to 155 Mbps via Prague Academic ATM connection in the beginning of 1996.
The computing capacity is coming to saturation and new ways of development are searched. In the HEP division of the Institute of Physics HP multiprocessor server usage is discussed. The Computing centre of the Institute plans to buy 2 SUN servers. The server in the HEP division of the Institute of Physics would serve mainly for batch processing for the all 4 laboratories. The group in the Faculty of Mathematics and Physics plans to add several workstations and X-terminals. The experimental data will be read from EXABYTE tapes to 20 GB buffer disk space. The intensive processing may be done either on the server nodes or even remotely when ATM connection becomes available to other laboratories. All laboratories can use nqs (nqs++) to submit batch jobs to the Institute of Physics or to their local stations. It is important to solve data access via DLT tapes and also system and user backup via automatic tools. The software is already available, as HEP FARM Prague already bought software site license from HP.
Table 
summarizes plans for development in the separate laboratories.
Table: Planned development of HEP FARM Prague.
The user friendly and user accessible backup system is crucial for such type of work.