Data set projections for Run12

The projected data set size for Run 12 was 600 TB of storage. The excess to an already 900 TB storage capacity by April 2012 was noted at the STAR analysis meeting (Apr16-20), BNL, NY in Software summary slide 13.

On 2012/05/14, the topic was brought again to the STAR trigger board. The summary is slightly off; what was reported is that

  • The issue is NOT a question of "limitations" but rather, where do we collectively stand - it is one thing to say "let's take all the data we can" but this requires planing S&C need to do but alos the RCF. Case and point, beyond the space we take on storage, the RCF need to buy a new HPSS silo as the current data samples exceeds by far the projected one (from last long term planing). This is significant.
  • The projected versus current reality was showed (numbers, Software summary, graph below)
  • The fact that our recent order of 300 tapes is all consumed already+an emergency order was placed for another 300 tapes ~ 450 TB.
  • With data samples of ~ PByte large and factors of (projected) x2 more than assumed AND the fact that our CPU resources are only 60% of the CPU needed, this implies that not only (as for Run 11) we will be able to only process one pass of the data in a year, but further, we will now (with the additional factor of x2 in data amount) NOT be able to make even one pass of production a year. STAR MUST be clear on this front ... "fight" for the datasets of interest will persist.

The figure below shows the status as per 2012/05/14

+ 800 TB of needed storage was projected as needed. With what's left, another 300 tapes will be ordered.

Note that the more data, the total purchase will need to be x2 for holding the DST (to date, we have purchased ~ 75k$ worth of tapes and will end up around a rough 200k$).