I think that there is some confusion here. The section that you refer to is not meant to address deduplication but rather TSMs native reclamation functionality. Reclamation is indeed used on tape hardware to consolidate data on tapes and thus to improve utilization. As you rightly point out, deduplication is another way that TSM can reduce backup target footprint, but that is a different topic that was not addressed in this post.

Thank you for your comment. Your point about throttling is absolutely right. The challenge is that there is still a minimum threshold and as drive speeds increase, this threshold gets higher. The other challenge is that hardware compression works against you since it reduces the amount of data written to tape and thus requires more bandwidth. (e.g. 2x compression reduces the effective bandwidth of data written to tape by 1/2.)

Managing tape is not all doom and gloom, but it certainly needs to be managed carefully and it is extremely difficult to achieve the state performance specifications. It often requires a trade-off between fastest backup (using lots of multiplexing) and fast restore speeds (limited multiplexing). In today’s data centers where data is more critical than ever, the choice of fast backup or fast restore is a difficult one to make.

Additionally: many backup programs use blocking factors which are too small for LTO (65kb or so). This needs to be cranked up to get best speed out of drives.

How about “personal”?

Georgens headed Engenio when they were part of Symbios Logic. When LSI bought Symbios they canned him.

Revenge is sweet, even though he is a moron.

Thank you for your comment. I think that we are in agreement on is the challenge. It is curious to see what NetApp does with Engenio’s controller technology.

Thank you for the follow-up. I believe that NetApp only purchased the external storage assets and thus the acquisition excluded things like MegaRAID. I believe that it also excluded ONStor for obvious reasons.