AMS Data Flow
AMS02 Data flow can be summarized in the following way: data are both recorded on board of the International Space Station (ISS) on AMS Laptop and directly transmitted, by means of satellites , to Marshall Space Flight Center (MSFC) in Alabama and therefrom, over the internet, to the Payload Operation Control Center (POCC). The CHD (critical health data) data are directly transmitted to the POCC in order to have an immediate overview of the detector status. The POCC is the control center of the AMS flight operations and where online data monitoring takes place to give a first look and evaluate the quality of data. From there, data will be sent to the Science Operation Center (SOC) where they will be processed and analyzed.
Finally data will be distributed from the SOC to Regional Sites that will act as storage of AMS Data samples and as Montecarlo production facilities to help the work of studying the detector response done by a simulation of real data at the SOC.
In this framework Milano DT works to transfer data from (to) CERN to (from) Regional Sites.
The Italian Ground Segment Data Storage (IGSDS), located at CNAF in Bologna (Italy), will host the master copy of the whole data sample of AMS. Both frames, raw, reconstructed and simulated data, expected to be around 150 TB per year, will be stored in IGSDS.
The IGSDS will act not only as a master copy, but also as a “distribution center”, in order to allow fast access to data to all collaborators without interfering with SOC activities.
IGSDS will also provide computing power for Montecarlo production and analysis facility for Italian collaborators.
AMS02 Milano DT
Data Transfer software was developed and is maintained by INFN Milano Bicocca since 2004. The main core of the Milano DT is a Multi-threaded finite state automa (written in Python) and the state transition jobs are written in Perl. It uses a database (Mysql/Oracle) for book-keeping and it can work in GRID & non-GRID environments.
Main goals of the AMS02 Milano DT is moving data efficiently from SOC to IGSDS without interfering with SOC activities and keeping track of what has been moved and if successfully or not. The minimum throughput required is at least 36 Mb/s which is three times the foreseen data rate so that, in case of DT malfunction it is possible to recover the fully alignment of the two copies (at CERN and at CNAF). The DT is also capable of moving data from remote locations to SOC in case of Montecarlo production in the regional sites.
In summary, thanks to flexibility and few prerequisites, the DT has been setup to/from the following locations:
IGSDS at CNAF, Bologna – Master copy of AMS02 Data Sample
ASI Science Data Center
Proof of concept and long duration stress tests OK
DT system is ready
Beijing Univ. Aeronautics and Astronautics (BUAA/NLAA)
South East University (SEU)