- Scott opined that backup solutions do not offer economies of scale, in his post [Backup Sucks Reason 38].
- I saw this as an opportunity to promote the new IBM Tivoli Storage Manager v6.1 which offers a variety of new scalability features, and continues to provide excellent economies of scale for large deployments, in my post [IBM has scalable backup solutions].
- Sadly, Scott decide to attack back with FUD, demonstrating that he has not been keeping up with the latest IBM TSM announcements, in his post [TSM Scalability (and the pot of gold at the end of the rainbow)].
"So does TSM scale? Sure! Just add more servers. But this is not an economy of scale. Nothing gets less expensive as the capacity grows. You get a more or less linear growth of costs that is directly correlated to the growth of primary storage capacity. (Technically, it costs will jump at regular and predictable intervals, by regular and predictable and equal amounts, as you add TSM servers to the infrastructure--but on average it is a direct linear growth. Assuming you are right sized right now, if you were to double your primary storage capacity, you would double the size of the TSM infrastructure, and double your associated costs.)"
I talked about inaccurate vendor FUD in my post [The murals in restaurants], and recently, I saw StorageBod's piece, [FUDdy Waters]. So what would "economies of scale" look like? Using Scott's own words:
- Without Economies of Scale
- "If it costs you $5 to backup a given amount of data, it probably costs you $50 to back up 10 times that amount of data, and $500 to back up 100 times that amount of data."
- With Economies of Scalee
- "If anybody can figure out how to get costs down to $40 for 10 times the amount of data, and $300 for 100 times the amount of data, they will have an irrefutable advantage over anybody that has not been able to leverage economies of scale."
- Case 1: The one-person company
- Here the sole owner needs a backup solution. Here are all the steps she might perform:
- Spend hours of time evaluating different backup products available, and make sure her operating system, file system and applications are supported
- Spend hours shopping for external media, this could be an external USB disk drive, optical DVD drive, or tape drive, and confirm it is supported by the selected backup software.
- Purchase the backup software, external drive, and if optical or tape, blank media cartridges.
- Spend time learning the product, purchase "Backup for Dummies" or similar book, and/or taking a training class.
- Install and configure the software
- Operate the software, or set it up to run automatically, and take the media offsite at the end of the day, and back each morning
- Case 2: The ten-person company
- I guess if each of the ten employees went off and performed all of the same steps as above, there would be no economies of scale.
Fortunately, co-workers are amazingly efficient in avoiding unnecessary work.
- Rather than have all ten people evaluate backup solutions, have one person do it. If everyone runs the same or similar operating system, file systems and applications, this can be done about the same as the one-person case.
- Ditto on the storage media. Why should 10 people go off and evaluate their own storage media. One person can do it for all ten people in about the same time as it takes for one person.
- Purchasing the software and hardware. Ok, here is where some costs may be linear, depending on your choices. Some software vendors give bulk discounts, so purchasing 10 seats of the same software could be less than 10 times the cost of one license. As for storage hardware, it might be possible to share drives and even media. Perhaps one or two storage systems can be shared by the entire team.
- For a lot of backup software, most of the work is in the initial set up, then it runs automatically afterwards. That is the case for TSM. You create a "dsm.opt" file, and it can list all of the include/exclude files and other rules and policies. Once the first person sets this up, they share it with their co-workers.
- Hopefully, if storage hardware was consolidated, such that you have fewer drives than people, you can probably have fewer people responsible for operations. For example, let's have the first five employees sharing one drive managed by Joe, and the second five employees sharing a second drive managed by Sally. Only two people need to spend time taking media offsite, bringing it back and so on.
- Case 3: The hundred-person company
- Again, it is possible that a hundred-person company consists of 10 departments of 10 people each, and they all follow the above approach independently, resulting in no economies of scale. But again, that is not likely.
- Here one or a few people can invest time to evaluate backup solutions. Certainly far less than 100 times the effort for a one-person company.
- Same with storage media. With 100 employees, you can now invest in a tape library with robotic automation.
- Purchase of software and hardware. Again, discounts will probably apply for large deployments. Purchasing 1 tape library for all one hundred people is less than 10 times the cost and effort of 10 departments all making independent purchases.
- With a hundred employees, you may have some differences in operating system, file systems and applications. Still, this might mean two to five versions of dsm.opt, and not 10 or 100 independent configurations.
- Operations is where the big savings happen. TSM has "progressive incremental backup" so it only backs up changed data. Other backup schemes involve taking period full backups which tie up the network and consume a lot of back end resources. In head-to-head comparisons between IBM Tivoli Storage Manager and Symantec's NetBackup, IBM TSM was shown to use significantly less network LAN bandwidth, less disk storage capacity, and fewer tape cartridges than NetBackup.
- The savings are even greater with data deduplication. Either using hardware, like IBM TS76750 ProtecTIER data deduplication solution, or software like the data deduplication capability built-in with IBM TSM v6.1, you can take advantage of the fact that 100 employees might have a lot of common data between them.
So, I have demonstrated how savings through economies of scale are achieved using IBM Tivoli Storage Manager. Adding one more person in each case is cheaper than the first person. The situation is not linear as Scott suggests. But what about larger deployments? IBM TS3500 Tape Library can hold one PB of data in only 10 square feet of data center floorspace. The IBM TS7650G gateway can manage up to 1 PB of disk, holding as much as 25 PB of backup copies. IT Analysts Tony Palmer, Brian Garrett and Lauren Whitehouse from Enterprise Strategy Group tried IBM TSM v6.1 out for themselves and wrote up a ["Lab Validation"] report. Here is an excerpt:
The majority of Fortune 1000 customers use IBM Tivoli Storage Manager, and it is the backup software that IBM uses itself in its own huge data centers, including the cloud computing facilities. In combination with IBM Tivoli FastBack for remote office/branch office (ROBO) situations, and complemented with point-in-time and disk mirroring hardware capabilities such as IBM FlashCopy, Metro Mirror, and Global Mirror, IBM Tivoli Storage Manager can be an effective, scalable part of a complete Unified Recovery Management solution.
"Backup/recovery software that embeds data reduction technology can address all three of these factors handily. IBM TSM 6.1 now has native deduplication capabilities built into its Extended Edition (EE) as a no-cost option. After data is written to the primary disk pool, a deduplication operation can be scheduled to eliminate redundancy at the sub-file level. Data deduplication, as its name implies, identifies and eliminates redundant data.
TSM 6.1 also includes features that optimize TSM scalability and manageability to meet increasingly demanding service levels resulting from relentless data growth. The move from a proprietary back-end database to IBM DB2 improves scalability, availability, and performance without adding complexity; the DB2 database is automatically maintained and managed by TSM. IBM upgraded the monitoring and reporting capabilities to near real-time and completely redesigned the dashboard that provides visibility into the system. TSM and TSM EE include these enhanced monitoring and reporting capabilities at no cost."
technorati tags: IBM, Tivoli, Storage Manager, TSM, TS7650, TS7650G, TS3500, Scalability, deduplication, economies+of+scale, Scott+Waterhouse, EMC, Symantec, NetBackup, backup, software, solutions, disk, tape, optical, drive