Re: VTFM vs TMM
Ron...thanks. I think this is possible only with hds. Thanks On Dec 1, 2010 3:34 AM, Ron Hawkins ron.hawkins1...@sbcglobal.net wrote: Techie, Your statement suggests that only internal drives have P-i-T IO consistency with replication products like TrueCopy and HUR. This is not correct. For virtualized midrange disks, synchronous and asynchronous replication is handled by the Enterprise Array providing the virtualization, not the midrange controller. There is no difference in IO consistency at the recovery site whether you use internal drives or virtualized midrange arrays for a TMM pool (or anything else). Ron -Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of techie well wisher Sent: Tuesday, November 30, 2010 10:56 AM To: IBM-MAIN@bama.ua.edu Subject: Re: [IBM-MAIN] VTFM vs TMM Tiering within the array is the best approach with replication. Synchronous or Asynchronous, Tiering within the array provides better consistent point at the recovery site. On Tue, Nov 23, 2010 at 11:49 AM, Ron Hawkins ron.hawkins1...@sbcglobal.net wrote: Techie well wisher, Thanks everyone. Much appreciated. With TMM, of course it goes to expensive z/os disk, but we do have the option of tiering within the array, such as using 1tb drives Raid6 and then hsm them to replicated vts later. [Ron Hawkins] That's not actually true. Cheaper midrange disk arrays can be virtualized by DASD controllers using a plethora of cheaper brands and models, and TMM can be tiered outside of the array without any appliance except the array itself. Internal SATA is an optional tier for those controllers that do not support virtualization of Mainframe volumes. Ron -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: VTFM vs TMM
Techie, True, but I was trying not to be so vendor centric. I like to think of it as something that can be done on CKD storage, but not by ever vendor. I think we all have products, features and functions that fit that description. Ron -Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of techie well wisher Sent: Friday, December 03, 2010 7:44 PM To: IBM-MAIN@bama.ua.edu Subject: Re: [IBM-MAIN] VTFM vs TMM Ron...thanks. I think this is possible only with hds. Thanks -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: VTFM vs TMM
Techie, Your statement suggests that only internal drives have P-i-T IO consistency with replication products like TrueCopy and HUR. This is not correct. For virtualized midrange disks, synchronous and asynchronous replication is handled by the Enterprise Array providing the virtualization, not the midrange controller. There is no difference in IO consistency at the recovery site whether you use internal drives or virtualized midrange arrays for a TMM pool (or anything else). Ron -Original Message- From: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] On Behalf Of techie well wisher Sent: Tuesday, November 30, 2010 10:56 AM To: IBM-MAIN@bama.ua.edu Subject: Re: [IBM-MAIN] VTFM vs TMM Tiering within the array is the best approach with replication. Synchronous or Asynchronous, Tiering within the array provides better consistent point at the recovery site. On Tue, Nov 23, 2010 at 11:49 AM, Ron Hawkins ron.hawkins1...@sbcglobal.net wrote: Techie well wisher, Thanks everyone. Much appreciated. With TMM, of course it goes to expensive z/os disk, but we do have the option of tiering within the array, such as using 1tb drives Raid6 and then hsm them to replicated vts later. [Ron Hawkins] That's not actually true. Cheaper midrange disk arrays can be virtualized by DASD controllers using a plethora of cheaper brands and models, and TMM can be tiered outside of the array without any appliance except the array itself. Internal SATA is an optional tier for those controllers that do not support virtualization of Mainframe volumes. Ron -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: VTFM vs TMM
Tiering within the array is the best approach with replication. Synchronous or Asynchronous, Tiering within the array provides better consistent point at the recovery site. On Tue, Nov 23, 2010 at 11:49 AM, Ron Hawkins ron.hawkins1...@sbcglobal.net wrote: Techie well wisher, Thanks everyone. Much appreciated. With TMM, of course it goes to expensive z/os disk, but we do have the option of tiering within the array, such as using 1tb drives Raid6 and then hsm them to replicated vts later. [Ron Hawkins] That's not actually true. Cheaper midrange disk arrays can be virtualized by DASD controllers using a plethora of cheaper brands and models, and TMM can be tiered outside of the array without any appliance except the array itself. Internal SATA is an optional tier for those controllers that do not support virtualization of Mainframe volumes. Ron -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: VTFM vs TMM
Maybe phrasing your question differently would be what are the advantages and disadvantages of physical tape data versus a tape-on- disk approach? TMM is a methodology, which temporarily stages tape data on Level 0 disk and then the TMM pool is managed by DFSMShsm or equivalent to consolidate this data on physical tape. The resulting physical tapes will then have data sets with varying expiration criteria and so will require recycling periodically and from a DR/BC viewpoint will require duplicating, as and if required. IBM Virtual Tape for Mainframe (VTFM) essentially is a virtual tape solution, emulating 3480/3490/3590 drives and allocating tape data to physical z/OS DASD. Of course, there are many other z/OS virtual tape solutions, with a tape-on-disk type concept, CA VTape being a software example, requiring physical tapes for data destaging, with Bus-Tech MDL/EMC DLm, Luminex, Universal Software, Intercom being appliance solutions that allocate tape data to FC/NAS disk arrays, without subsequent data destaging to physical tape, and of course IBM TS7700 (VTS), Oracle/StorageTek VSM and FSC CentricStor being solutions that combine a disk cache and physical tape. So thinking of Sherlock Holmes when you have eliminated the impossible, whatever remains, however improbable, must be the truth, maybe you could review all of the tape-on-disk options, which include TMM? Some advantages of those solutions, and so for the avoidance of doubt, Bus-Tech MDL/EMC DLm, Luminex, Universal Software, Intercom, et al, is that the resulting ML2 type data, can be easily recycled, as the tape data is on cost-efficient FC/SAN disk, and thus easier data replication for BC/DR is also possible. Equally, ML1 type operations could also be eliminated, with all of the resource considerations (E.g. CPU, z/OS class DASD) associated with that process. Thus for the avoidance of doubt, avoid ML1 disk costs and zSeries CPU cycles, by eliminating ML1 from the storage hierarchy and go direct to ML2, where compression is performed outboard of the Mainframe and tape data allocated on less expensive FC/IP disk arrays, potentially with the benefits of deduplication. All that said, maybe even TMM can co-exist with such a tape-on-disk methodology. As with any IT solution, identify your business requirements first and then research what products best fit your business requirements with the best ROI and TCO attributes. So maybe VTFM isnt for you and maybe TMM can be approached from a different viewpoint for you by utilizing other virtual tape technologies. On Mon, 22 Nov 2010 11:30:08 -0500, techie well wisher techi...@gmail.com wrote: IBM has VTFM (which is diligent/copycross, etc). Why do we need this product or use this product while we can directly intercept and direct allocations to a particular storage group (such TMMGROUP) with disk volumes, let's say a dedicated set aside pool from a storage device? With extended dataclas attribute, the datasets in this group could be really huge (several gigabytes). To me, this product adds unnecessary complexity. With this, we don't need PAT (parallel tape access), because all the datasets in this group are disk datasets, accessible by multiple address spaces/jobs. Pleaset let me know your thoughts or am I missing something here? TW -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: VTFM vs TMM
Thanks everyone. Much appreciated. With TMM, of course it goes to expensive z/os disk, but we do have the option of tiering within the array, such as using 1tb drives Raid6 and then hsm them to replicated vts later. On Nov 23, 2010 4:39 AM, Michael W. Moss mikey.m...@virgin.net wrote: Maybe phrasing your question differently would be what are the advantages and disadvantages of physical tape data versus a tape-on- disk approach? TMM is a methodology, which temporarily stages tape data on Level 0 disk and then the TMM pool is managed by DFSMShsm or equivalent to consolidate this data on physical tape. The resulting physical tapes will then have data sets with varying expiration criteria and so will require recycling periodically and from a DR/BC viewpoint will require duplicating, as and if required. IBM Virtual Tape for Mainframe (VTFM) essentially is a virtual tape solution, emulating 3480/3490/3590 drives and allocating tape data to physical z/OS DASD. Of course, there are many other z/OS virtual tape solutions, with a tape-on-disk type concept, CA VTape being a software example, requiring physical tapes for data destaging, with Bus-Tech MDL/EMC DLm, Luminex, Universal Software, Intercom being appliance solutions that allocate tape data to FC/NAS disk arrays, without subsequent data destaging to physical tape, and of course IBM TS7700 (VTS), Oracle/StorageTek VSM and FSC CentricStor being solutions that combine a disk cache and physical tape. So thinking of Sherlock Holmes “when you have eliminated the impossible, whatever remains, however improbable, must be the truth”, maybe you could review all of the tape-on-disk options, which include TMM? Some advantages of those solutions, and so for the avoidance of doubt, Bus-Tech MDL/EMC DLm, Luminex, Universal Software, Intercom, et al, is that the resulting ML2 type data, can be easily recycled, as the “tape” data is on cost-efficient FC/SAN disk, and thus easier data replication for BC/DR is also possible. Equally, ML1 type operations could also be eliminated, with all of the resource considerations (E.g. CPU, z/OS class DASD) associated with that process. Thus for the avoidance of doubt, avoid ML1 disk costs and zSeries CPU cycles, by eliminating ML1 from the storage hierarchy and go direct to ML2, where compression is performed outboard of the Mainframe and tape data allocated on less expensive FC/IP disk arrays, potentially with the benefits of deduplication. All that said, maybe even TMM can co-exist with such a tape-on-disk methodology. As with any IT solution, identify your business requirements first and then research what products best fit your business requirements with the best ROI and TCO attributes. So maybe VTFM isn’t for you and maybe TMM can be approached from a different viewpoint for you by utilizing other virtual tape technologies. On Mon, 22 Nov 2010 11:30:08 -0500, techie well wisher techi...@gmail.com wrote: IBM has VTFM (which is diligent/copycross, etc). Why do we need this product or use this product while we can directly intercept and direct allocations to a particular storage group (such TMMGROUP) with disk volumes, let's say a dedicated set aside pool from a storage device? With extended dataclas attribute, the datasets in this group could be really huge (several gigabytes). To me, this product adds unnecessary complexity. With this, we don't need PAT (parallel tape access), because all the datasets in this group are disk datasets, accessible by multiple address spaces/jobs. Pleaset let me know your thoughts or am I missing something here? TW -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: VTFM vs TMM
Techie well wisher, Thanks everyone. Much appreciated. With TMM, of course it goes to expensive z/os disk, but we do have the option of tiering within the array, such as using 1tb drives Raid6 and then hsm them to replicated vts later. [Ron Hawkins] That's not actually true. Cheaper midrange disk arrays can be virtualized by DASD controllers using a plethora of cheaper brands and models, and TMM can be tiered outside of the array without any appliance except the array itself. Internal SATA is an optional tier for those controllers that do not support virtualization of Mainframe volumes. Ron -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
VTFM vs TMM
IBM has VTFM (which is diligent/copycross, etc). Why do we need this product or use this product while we can directly intercept and direct allocations to a particular storage group (such TMMGROUP) with disk volumes, let's say a dedicated set aside pool from a storage device? With extended dataclas attribute, the datasets in this group could be really huge (several gigabytes). To me, this product adds unnecessary complexity. With this, we don't need PAT (parallel tape access), because all the datasets in this group are disk datasets, accessible by multiple address spaces/jobs. Pleaset let me know your thoughts or am I missing something here? TW -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: VTFM vs TMM
W dniu 2010-11-22 17:30, techie well wisher pisze: IBM has VTFM (which is diligent/copycross, etc). Why do we need this product or use this product while we can directly intercept and direct allocations to a particular storage group (such TMMGROUP) with disk volumes, let's say a dedicated set aside pool from a storage device? With extended dataclas attribute, the datasets in this group could be really huge (several gigabytes). To me, this product adds unnecessary complexity. With this, we don't need PAT (parallel tape access), because all the datasets in this group are disk datasets, accessible by multiple address spaces/jobs. Pleaset let me know your thoughts or am I missing something here? Well... Diligent is emulated tape drive, data physically reside on your DASD. Potentially the most expensive DASD in your shop. Advantages: many tape devices (consurrent jobs), simple replication, fast tape mounts. Disadvantages: occupies disk, CPU consuming, especially with compression (zIIP could relieve it), space is constrained by size of your DASD. TMM is disk cache backed with real tapes. Advantages: it's free, it fills up (tries to) your tape volumes, quite easy to set up. Disadvantages: it may not solve all your needs. My humble opinion: it's good idea to get rid off the tapes in data processing with the exception for activities like ML2, backups, dumps and archive. YMMV, but usually we want to have data on DIRECT ACCESS Storage Device. Direct access is good. -- Radoslaw Skorupka Lodz, Poland -- BRE Bank SA ul. Senatorska 18 00-950 Warszawa www.brebank.pl Sd Rejonowy dla m. st. Warszawy XII Wydzia Gospodarczy Krajowego Rejestru Sdowego, nr rejestru przedsibiorców KRS 025237 NIP: 526-021-50-88 Wedug stanu na dzie 16.07.2010 r. kapita zakadowy BRE Banku SA (w caoci wpacony) wynosi 168.248.328 zotych. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
Re: VTFM vs TMM
Radoslaw, Disclaimer: I am a TMM bigot, and I agree 110% (one hundred and ten percent) with the OP. TMM is disk cache backed with real tapes. [Ron Hawkins] I don't quite agree with your definition. TMM can be, and often is migrated and stacked onto to tape but the disk caching does not need to be backed by tape storage. Datasets can live and die in a TMM pool without ever being migrated. I also believe that Management Class provides a greater set of policies for managing dataset residency in the TMM pool than you get with Virtual tape systems. Advantages: it's free, it fills up (tries to) your tape volumes, quite easy to set up. [Ron Hawkins] What about: there is always zero mount time in the TMM Pool; there is no limit on number of drives active; Many concurrent readers; Generally faster than TMM appliances; DSORG=PS datasets restored with FCV2, FCV2 manual migration to cheap disk; P-i-T synchronization with DASD for DR Disadvantages: it may not solve all your needs. [Ron Hawkins] What about: Staging from ML2, No Deduplication, FCV2 manual migration to STORGRUPS on cheap disk; ML2 Tape needs DUPLEX with remote vaults for DR (or similar). My humble opinion: it's good idea to get rid off the tapes in data processing with the exception for activities like ML2, backups, dumps and archive. YMMV, but usually we want to have data on DIRECT ACCESS Storage Device. Direct access is good. [Ron Hawkins] Agreed Ron -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
