Originally posted by: TonyPearson

Normally, IBM only makes announcements on Tuesdays, but today, Friday, IBM announces that it acquired Diligent Technologies. What? I got a lot ofquestions about this, so I thought I would start with this...

[YouTube clip montage of "LOST"] from [my friend Pam]

When I posted in January that[IBM Acquires XIV],fellow EMC blogger Mark Twomey of StorageZilla fame, sent me a comment:

"Ah now Tony I wasn't poking fun. Indeed I find it fascinating that Moshe who's been sitting out on the fringes for years having been banished for being an obstructionist to EMC entering the mid-market is now back.

Which reminds me what happens with Diligent? There his as well aren't they or has he packed his stake in that in?"

As you might have guessed, I am privy to a lot of stuff going on behind the scenes at IBM that I can't talk about in this blog, and all these rumors in the blogosphere about IBM acquisition of Diligent was a topic I couldn't officially recognize, defend or deny, until official IBM announcements were made.

In his latest post, Mark wonders about[the last Tape and Mainframe sales person on earth]. He recounts my interaction with fellow HDS blogger Hu Yoshia about the energy benefits ofVirtual Tape Libraries. Knowing that we were going to announcement IBM's acquisition of Diligent soon, I thoughtthis would be a worthy exchange, driving up the sales of Diligent boxes (whether you buy them from IBM or HDS).Diligent already had reselling arrangements with HDS, and IBM plans to continue thosearrangements going forward with HDS. As I have explained before in my post [Supermarketsand Specialty Shops], IBM and HDS cater to different customers, so if a customer who wants the best technologyfrom a specialty shop, they can buy IBM Diligent products from HDS, but if they want one-stop shopping, they can buyIBM Diligent directly from IBM or its other IBM Business Partners.

(Perhaps a more tricky situation is that Diligent also had an arrangement with Sun Microsystems, which competesdirectly against IBM as another IT supermarket vendor, but I have not heard how IBM has decided to handle thisgoing forward.)

For more on this intricate mess of interconnected companies, alliances and partnerships, read Dave Raffo's article[Data dedupe dance cardfilling up] over at Storage Soup.

So, let's tackle the first question:

Q1. What will happen to IBM's real tape library business?

Come on! IBM is Number one in tape, we've had virtual tape libraries since 1997 (the first in the industry)and continue to do well in both virtual and real tape libraries. Both provide value to the customer, and bothhave their place as part of the overall "information infrastructure". This acquisition provides yet another choicefor clients on our "supermarket" shelf.

(For those following the ["which is greener"] discussion, the robot of the IBM TS3500 real tape library consumes185W per frame (when moving) and each tape drive consumes 50W (when actively working on a tape). Compared to 13W per SATA disk drive, each 6-drive frame of a TS3500 consumes as much electricity as 37 SATA disk drives. If you are not running backups 24x7, the total KWh per day for your tape library is actually quite less, but as several people have pointed out, there are customers that do run backups 80-90 percent of the time. LTO-4 tapes can hold 800GB uncompressed, and SATA disk are now available in 1TB (1000 GB) size, so you can have fun with your own comparisons.)

Meanwhile, Scott Waterhouse, one of the few people at EMC who understand tape workloadslike backup and archive, takes me to task in his Backup Blog with his post[I want a Red Ferrari].For those who are surprised that anyone at EMC might understand backup workloads, EMC did acquire a company calledLegato, and perhaps Scott came from that acquisition. I've never met Scott in person, but based solely only fromhis writings, he seems to know his stuff and makes strong arguments for using IBM Tivoli Storage Manager (TSM) with deduplication and virtual tape libraries.

While TSM does a good job of "deduplicating" at the client first, backing up only changed data, Scott feels database and email repositories must be backed up entirely each time, which is what happens in many other backup software products. Some clients might have 80 percent database/email and only 20 percent files, while others might have less than 20 percent database/email and 80 percent files, so this might influence whether deduplication will have small or big benefit.If TSM has to backup the entire database, even though little has changed since the last backup, that is where deduplication on a virtual tape library can come in handy. For IBM DB2 and Oracle databases, IBM TSM application-aware Tivoli Data Protection module interface backs up only changed data, not the entire file. Thanks to IBM's FilesX acquisition-- (also coincidently from Israel) --IBM can extend this support now to SQL Server databases as well.However, to be fair, Scott is partly correct, TSM does backup some database and email repositories in their entirety, which is why it is a good idea to have BOTH an IBM virtual tape library with deduplication and Tivoli Storage Manager to handle all cases. This brings us to the next question:

Q2. What will happen to IBM's patented "progressive backup" technology?

IBM will continue to use TSM's progressive backup technology. TSM already works great with Diligent virtual tapelibraries. One example is LAN-free backup. In this configuration, the TSM client writes its backups directly toa virtual or real tape library, over the SAN, and then sends the list of files backed up to the TSM server over theLAN to record in its database. This can greatly reduce IP traffic on your LAN during peak backup periods. For more about this, see the IBM Redbook titled["Get More Out of Your SAN with IBM Tivoli Storage Manager"].

Jon Toigo from DrunkenData asks[Did IBM Do Due Diligence Before Making Diligent Acquisition a Done Deal?] which is probably always a valid question. Unlike XIV, I wasn't part of the Diligent acquisition team, so I can't provide first hand account of the process. I am told that the IBM team did all the right things to make sure everything is going to turn out right.Sadly, many companies that make acquisitions in the IT industry fail to make them work. Fortunately, IBM is one of the few companies that has a great success record, with over 60 acquisitions in the past six years.In the Xconomy forum, Wade Rousch writes[IBM and the Art of Acquisitions]and gives some insight why IBM is different. Jon did not understand why Cindy Grossman, IBM VP of tape and archive solutions, ran the analyst conference call for this announcement, which brings me to the next question:

Q3. What is Diligent virtual tape library going to be categorized as, a disk system or a tape system?

IBM organizes its storage systems based on the host application workloads.Products to address disk workloads (SVC, DS8000 series, DS6000 series, DS4000 series, DS3000 series, N series, XIV Nextra) are in our disk systems group. Storage that appears to host applications like a tape system to address workloads like backup and archive (tape drives, libraries and tape virtualization) are in our tape and archive group. IBM Diligent has two products, one for big workloads and one for medium workloads. Both look liketape systems, so our tape and archive team, who understand tape workloads like backup and archive the best, are obviously the best choice to support IBM Diligent in the mix.

Some might remember that IBM already has deduplication in its IBM N series products, called Advanced Single InstanceStorage, or A-SIS. Here is an[IBM System Storage N Series A-SIS Deduplication Deployment and Implementation Guide. This brings up the fourth question:

Q4. What will happen to N series A-SIS feature?

IBM will offer both N series and Diligent deduplication capabilities. For disk workloads, IBM N series offers a post-process deduplication feature at no additional charge. For tape workloads, IBM will now offer an in-line deduplication feature with Diligent Technologies. Different workloads, different offerings.

As with any acquisition, there will be some changes. The 100 folks from Diligent will get to learn the IBM wayof doing things. This brings me to our fifth and final question:

Q5. What is the correct spelling: deduplication or de-duplication?

It appears that Diligent has a corporate-wide standard to hyphenate this term (de-duplication), but the "word police" at IBM that control and standardize all "proper spellings, trademarks, and capitalization" have sent me corporate instructions a few days ago that IBM does not to hyphenate this term (deduplication). So, going forward, it will be "deduplication", or "dedupe" for short.I suspect one of the first tasks that our new IBMers from Diligent will be doing is removing all those hyphens fromthe [Diligent Technologies website]!

That's all for now, I'm off to Chicago, Illinois tomorrow!

technorati tags: IBM, Diligent, acquisition, LOST, EMC, Mark Twomey, Moshe Yanai, Pam West, StorageZilla, HDS, Hu Yoshida, Sun Microsystems, virtual, real, tape library, VTL, SATA, LTO, LTO-4, TS3500, Scott Waterhouse, information infrastructure, email, database, backup, archive, Oracle, DB2, SQL Server, FilesX, SAN, LAN, IP, LAN-free, TSM, Tivoli Storage Manager, XIV, Nextra, A-SIS, Chicago, deduplication, Dave Raffo, dedupe