IBM supplies a perl script to measure the cost of dedup. See http://www-01.ibm.com/support/docview.wss?uid=swg21596944
I just ran it in an instance with an 800 GB db, here are the final summary lines - ---------------------------------------- Final Dedup and Database Impact Report ---------------------------------------- Deduplication Database Totals ----------------------------- Total Dedup Chunks in DB : 1171344436 Average Dedup Chunk Size : 447243.5 Deduplication Impact to Database and Storage Pools --------------------------------------------------- Estimated DB Cost of Deduplication: 796.51 GB Total Storage Pool Savings: 230466.30 GB That works out to ~3.5 GB per TB saved. The db is not on SSD. It is on a 6 disk raid 10 array internal on a Dell server. Overall I am very happy with TSM dedup. Thanks, Bill Colwell Draper lab -----Original Message----- From: ADSM: Dist Stor Manager [mailto:ADSM-L@VM.MARIST.EDU] On Behalf Of Dan Haufer Sent: Thursday, June 12, 2014 4:31 PM To: ADSM-L@VM.MARIST.EDU Subject: Re: TSM and VTL Deduplication Yes, one of the two. If TSM deduplication is enabled and the target is a virtual tape, i doubt if the VTL can deduplicate anything from the write data. -------------------------------------------- On Thu, 6/12/14, Ehresman,David E. <deehr...@louisville.edu> wrote: Subject: Re: [ADSM-L] TSM and VTL Deduplication To: ADSM-L@VM.MARIST.EDU Date: Thursday, June 12, 2014, 12:51 PM Unless you have a specific requirement, I would suggest you choose either TSM dedup to disk or go straight to virtual tape. There is not usually a need to do both. David -----Original Message----- From: ADSM: Dist Stor Manager [mailto:ADSM-L@VM.MARIST.EDU] On Behalf Of Dan Haufer Sent: Thursday, June 12, 2014 2:41 PM To: ADSM-L@VM.MARIST.EDU Subject: Re: [ADSM-L] TSM and VTL Deduplication Thanks for all the answers. So SSDs (Looking at SSD caching) for the database storage and 10GB per TB of total backup data on the safer side. -------------------------------------------- On Thu, 6/12/14, Erwann Simon <erwann.si...@free.fr> wrote: Subject: Re: [ADSM-L] TSM and VTL Deduplication To: ADSM-L@VM.MARIST.EDU Date: Thursday, June 12, 2014, 8:47 AM Hi, I'd rather say 6 to 10 times, or 10 GB of DB for each 1 TB of data (native, not deduped) stored. -- Best regards / Cordialement / مع تحياتي Erwann SIMON ----- Mail original ----- De: "Norman Gee" <norman....@lc.ca.gov> À: ADSM-L@VM.MARIST.EDU Envoyé: Jeudi 12 Juin 2014 16:55:29 Objet: Re: [ADSM-L] TSM and VTL Deduplication Be prepare for your database size to double or triple if you are using TSM deduplication. -----Original Message----- From: ADSM: Dist Stor Manager [mailto:ADSM-L@VM.MARIST.EDU] On Behalf Of Prather, Wanda Sent: Thursday, June 12, 2014 7:15 AM To: ADSM-L@VM.MARIST.EDU Subject: Re: TSM and VTL Deduplication And if you are on the licensing-by-TB model, when it gets un-deduped (reduped, rehydrated, whatever), your costs go up! -----Original Message----- From: ADSM: Dist Stor Manager [mailto:ADSM-L@VM.MARIST.EDU] On Behalf Of Dan Haufer Sent: Thursday, June 12, 2014 9:48 AM To: ADSM-L@VM.MARIST.EDU Subject: Re: [ADSM-L] TSM and VTL Deduplication Understood. Thanks ! -------------------------------------------- On Thu, 6/12/14, Ehresman,David E. <deehr...@louisville.edu> wrote: Subject: Re: [ADSM-L] TSM and VTL Deduplication To: ADSM-L@VM.MARIST.EDU Date: Thursday, June 12, 2014, 5:33 AM If TSM moves data from a (disk) dedup pool to tape, TSM has to un-dedup the data as it reads it