COMPACTION only affects active data sets. GFTAVMA models backup data sets as migrating directly to migration level 2 and to whatever device you specify with the TAPEDEV and TAPELEN keywords.

DFSMShsm compaction is taken into account for data sets that GFTAVMA determines to belong on migration level 2 tape based on the length of time since they have been referenced. These data sets would have first resided on migration level 1 for some period of time. For these data sets, GFTAVMA applies the nn percentage compaction that would have been achieved during migration to migration level 1 and then assumes an additional 15% IDRC compaction during the level 1 to level 2 migration.

Data sets that migrate directly from primary storage to migration level 2 do not require DFSMShsm software compaction if IDRC is available to the tape unit controllers (3480/3490). This saves processor and IDRC compaction at the bit level and is more efficient than the software compaction at the byte level. Data sets that are already software compacted can usually only be IDRC compacted an additional 10%.

Therefore, when a 600MB data set with a specified IDRCFACTOR of 3.0 migrates from primary DASD to migration level 2 tape, the data set is compacted to 200MB. If the same 600MB data set has COMPACTION(50) and IDRCFACTOR(3.0), it is compacted 50% to 300MB when it migrates to migration level 1. When it migrates to migration level 2, it is reduced an additional 45MB to 255MB. Thus, the IDRC compacted data set achieves a total of 200MB, while the software compacted and IDRC compacted data set achieves only a total of 255MB.

Therefore, you must consider the effects of using DFSMShsm's software compaction with IDRC compaction. A data set that has been software compacted cannot achieve full IDRC compaction.