Login Form

About Me

RSS Feed

Mr. Backup Mr. Backup

More misinformation about backups

PDFPrintEmail

I don't care if you use disk, tape, or the cloud to back up your systems.  (In case you think I'm swayed by advertising, I have advertisers from all of those categories.)

Having said that, it bothers me when I see misinformation being used to sway you one way or the other.  This is why I wrote this article that disproved the Gartner 71% tape failure "quote," and this article disproving the Yankee Group 42% failure "quote."  And since he used my comment system to link to his article, I also thought I'd write this blog article dispelling the misinformation in the article he linked to.

He said it's been a long time since people have seen tape used for backups. 

The live survey of the hundreds of attendees to last year's Backup Central Live shows showed that 82% of them still use tape as their final destination for backups.  So much for not seeing tape in a while.

He said IT pros are still skeptical that removable drives have a legitimate place in backup

Yes, we are.  I think that 3.5" removable disk drives are a very bad place for backups. 2.5" drives yes, 3.5" drives, not so much.  They're simply not designed for excessive portability.  Adding to that is this fact:  every portable hard drive I have ever used for backing up my laptop has died long before the drive it was backing up.  Every single one.

He said cloud backup is shiny and new and that's why people are choosing it.

No, it's because it's a complete and total outsourcing of backup functionality.  Backups can be onsite and offsite without ever touching a disk drive or tape drive.  AND you will be constantly notified if your backups are working or not working.  You often even get notified even if you shut off all your backups!  That's not the case with any backup software product that I've ever used.  There are a lot of reasons to use cloud backup over removable disk drives or tape.  In fact, there are so many that I strongly recommend cloud backup for small to medium sized companies.

He said disk is cheaper for small companies

Yes, it is.  It is cheaper to acquire the drives as long as you never need to add capacity.  If you do, however, need to add capacity, disk costs will double.  Tape costs will not.  It'll cost you about $.02/GB to add more capacity to a tape-based system.  (Having said that, I do not recommend backing up directly to tape; I haven't in a while.)  Having said that, I priced a slightly different tape-based system than the one he quoted in his article, and it was approximately the same cost.

As to comparing their 10-bay disk systems against an autoloader, I don't see how you can do that.  Backup software products simply don't know what to do with 10 removable disk drives, but they do not what to do with autoloaders.  (I'm sure he knows a backup software product that will work his configuration, but I don't know of one.)

He said disk is more reliable than tape

Baloney.  I've written about this before.  Tape has a much higher reliability rate than SATA disk -- one hundred times more reliable.  I've already shown above that the statistics he quotes in his article are bunk.  Almost every failure I've ever seen with restores was the fault of anything but the media that was being used.  (I still don't think tape should be used as the initial target for backups, but it is a very reliable place to put the second copy.)

He said LTO-5 speed is 140 MB/s, but only with compression

Sorry, Charlie.  That's the native speed.  It's up to 280 MB/s with compression.  With the 1.5:1 I see all the time, it's 210 MB/s easy.  Having said that, it's very hard to feed data to a drive that fast.  This is why I don't recommend tape as the initial target for backups.  But if you've already made a copy on disk, you should have no problem streaming that tape drive.

He said single file restores are faster from disk

Yes, they are.  That extra minute it takes the tape to load and get to the single file would probably put most companies out of business.  Seriously, you're going to make a case out of a minute of tape loading time?

He said upgrading/replacing is cheaper with disk

Are you kidding me?  I addressed this already.  If you need more capacity than the initial purchase, it's much cheaper to expand a tape system. Most customers go years without upgrading their drives.

He said doing synthetic fulls is easier on disk

Yes, it is.  And CDP and CDP-like tech is also only possible on disk.  Disk has a lot of things going for it.

He said tape has to be replaced more often

Again, baloney.  A tape that is used once a week will last four years with the chart quoted in the article.  That is longer than most disks I've used.

Summary

I don't recommend using tape as the initial target for backups, but I still think it's a great place to put the next copy.  And if you are a smaller company, I think the best thing you can do is to use a cloud backup service that totally automates everything, and alllows for a local copy of your data.  But I think that any backup system that requires small companies to manually swap removable media just to make backups happen is a bad idea.

Comments   

 
0 #13 W. Curtis Preston 2012-06-28 15:57
Any email from us has a valid return address, our business address, and a one-step opt-out link. If you choose not to use that opt-out link, responding to the email will give you instructions on how/where to send a manual unsubscription request.
 
 
0 #12 please stop the spam 2012-06-28 14:45
I'm getting spammed by email purportedly from you, with fake or broken return address. How to stop the spam from you??
 
 
0 #11 Darren McBride 2012-03-29 17:08
When I googled for tapes I thought I found them for $51 each wholesale online not $75. $51/1500 is 3.4 cents per gig. Curtis disagrees with me both because LTO-4 tapes are cheaper and because he uses 2 to 1 compression. Your experience validates my point but it is unusual to only get 1.2 to 1
 
 
0 #10 Avram Woroch 2012-03-29 06:46
A question about the $0.02/GB quoted - may I ask how that is coming up? I ask because we're using LTO5 @ $75 CAD each, and we're getting about 1.2:1 with compression. I figure our costs are easily double that.
 
 
0 #9 Darren McBride 2012-03-19 02:43
Yes tape vendors do the math that way and I understand their compression chips can help throughput, I just don’t think we should let them get away with it. Do Dedupe vendors advertise say a 1TB appliance as 10TB or 20TB because most people will have 10x or 20x deduplication? Certainly they MENTION this but they wouldn’t have a headline saying “20TB dedupe appliance” if it only had a 1TB drive in it. Whether the compression slows the backup down slightly or speeds it up (your explanation is consistent with my tests & makes sense), the fact remains that compression is available and works well in most popular imaging software (acronis, symantec System restore, shadowprotect, etc) and is turned on by default when you take an image. I wouldn’t be surprised if Intel’s new AES-NI capable CPUs start making software compression really fly, possibly making backups to disk faster with compression on than off (1/2 the data moves to the disk). If that happens I guess it’d be OK to begin quoting 2 times the capacity on our drives? I have two choices when quoting backup capacities of removable hard drive backup. 1) Do what tape vendors do and tell customers that a 1.5TB hard drive backs up about 3TB 2) Quote the actual 1.5TB drive size and then point out the other guys are fudging the numbers.

I understand what you’re saying about backup and replication software being integrated. I believe that many packages do provide both, especially cloud vendors. I’ve noticed that the backup vendors do backup well but not so well at replication and cloud vendors are good at replicating but don’t backup well. What I mean is some of the low cost cloud vendors you mention don’t support backup or granular (message level) restore of Exchange and don’t do enterprise products like AD, SQL, or Sharepoint well. The software I use a lot is ShadowProtect. It is a backup/ drive imaging product but they charge extra to replicate offsite so that was my point of reference. They have an FTP product and a faster even more expensive “Shadowstream” product which requires software on both ends . Some of their competition (appasure is one I can think of) includes replication as do the SAN and high end vendors. In fact, now that I think about it, StorageCraft might be one of the few that don’t include replication but I think Symantec System Restore is another popular one that doesn’t.

I think your points about the extra (sometimes hidden) labor cost of handling tapes or drives is valid. Also true that is they are more subject to human error. As far as paying a vaulting company I’ve noticed very few small business clients (50 seats and 5 servers or less is my point of reference) pay an Iron Mountain (or anyone) to take media off site, relying instead on an employee to take it home (Yes and probably regularly forget it & leave it in their car overnight to freeze). In fact many of them argue that simply leaving media onsite in a fireproof safe is good enough (I don’t agree).

I’m also going to agree with you that backup experts should be recommending that customers backup to an on-site appliance or drive FIRST, and then back that up to cloud, tape, or removable drive (secondary backup and archive). 90% of your restores would then be from the fixed unit. From my experience the response to this recommendation is often “we can barely afford one backup device – let alone two”. I think our difference of opinion is that for secondary and offsite I would prefer cloud and/or removable drive whereas you would use cloud and/or tape .

Believe it or not, we do have a few clients that backup small servers without using backup software (softwareless backup), although most of them prefer to do it the conventional way. The way they do it is to share up one drive of a 2 bay mirroring unit to their end users, and pull the other drive out nightly. After inserting a new removable drive the system automatically remirrors in hardware overnight and keeps the replica up to date in real time the next day. It can even be done by mirroring RAID5 using hardware to a stand alone removable drive. The main disadvantages are the speed of the shared volume (versus server class RAID drives) and the 3TB size limit. You’ll undoubtedly point out others (like relying on 1 device for both primary and backup storage). But a sub $1000 backup system can be compelling.
 
 
0 #8 W. Curtis Preston 2012-03-18 20:18
Everyone but you seems OK with doing the math with some sort of compression, which brings the tape to less than $.02/GB.

I know, you don't think that's fair and that you should just do the compression in the backup software. The only problem with that is that it has always taken longer to do compression in the backup software than to do it in a chip in the tape drive. The chip in the tape drive makes the tape drive faster; doing it in backup software makes backups and restores slower -- in all but a few edge cases. So this is why we use compression numbers in tape drives. 1) They're real. Most all open-systems data compresses at a rate of at least 1.5:1 and often better than that. 2) It's easier and better to do it in the device.

I think your pricing discussion on cloud is a bit clouded. ;) You're not buying backup software and replication software. You're buying backup software that can replicate. And are you implying that if a customer uses your hardware, they don't have to buy backup software? And yes, a cloud customer has to pay for bandwidth. A person using disk or tape for offsiting without replication has to pay for that service. Nothing is free.

BUT if you use intelligent enough backup software that does dedupe and you get a good enough dedupe rate, then you often don't have to buy extra bandwidth for your backups; they can just use the bandwidth that you paid for during the day, but that is not used much at night. So SOME customers (I don't have a % as to how many) aren't paying for additional bandwidth. Cloud customers have their data stored at a cloud company; your customers have their data stored at a vaulting company. We all have to trust someone.

Regarding tape revenue. You hit the nail on the head; it has indeed been decreasing for a number of reasons. The first reason is the huge move of people to disk as their initial target for backups. (My own research shows that 82.6% of users I've surveyed still use tape as their ultimate destination for backups.) The less obvious reason is that while people are still buying more drives and tapes than they've ever bought before, those tapes are storing more and more data and are costing less and less, resulting in lower revenues per cartridge.

Regarding your Maxell tape analyzer... Most modern tape libraries have that functionality built-in. It's the eqivalent of SMART for disk drives.
 
 
0 #7 Darren McBride 2012-03-18 18:54
$51 LTO-5 tape divided by 1500GB is 3.4 cents per Gig, not 2 cents.
I ran across this interesting device by Maxell for analyzing LTO tapes. I think it'd be a good thing to have for anyone using lots of them. http://www.tape4backup.com/183100.php#
Curtis, I officially apologize for implying your opinion might have been based on sponsorship. Certainly I cannot claim to be objective myself but can say my opinion is also based on experience. I linked to those articles saying explicitly they've been dis proven by You, which I thought gave your expertise a plug but I will take them out. I agree having a local drive or appliance and THEN off-siting to cloud is the right way to do cloud in a business environment, but it does mean you're buying the local storage, plus backup software, plus replication software, plus Internet bandwidth, plus paying the vaulting fees. (and assumes you're comfortable with your data being somewhere off-site).

New tape numbers are out showing slow decline in tape sales, but with small bump expected this year as LTO-5 comes online. 90% of tape (media) sales are LTO now. Sales of tapes have declined over last 4 years from 1,255 Million in 2008 to 771 million in 2011.
21% of units being sold are LTO-5 (with 18% growth) and 46% of units are LTO-4.
so it seems comparing prices to LTO-5 wasn't completely crazy. http://www.storagenewsletter.com/news/marketreport/sccg-4q11-tape
 
 
0 #6 W. Curtis Preston 2012-03-13 20:50
@Joe

This particular blog post was a response to another person's blog post. That blog post is from someone who is marketing standalone, portable SATA drives as a replacement for tape, under the guise that they are unreliable. My knowledge of their architecture and my almost 20 years experience in both devices says otherwise.

Having said that, the fact that you can put disk in a RAID set and protect it against individual media failure is one of its biggest advantages.

I did want to say something about the "I don't know a tape is bad till I need it" comment. First, the same would be true of any removable medium. Second, if you have the budget to store that offsite tape in a tape library that supports automated media testing (e.g. Quantum or Spectralogic) then that wouldn't have to be the case. They can do a full scan of every block and compare it against that block's CRC that is also stored on the tape.
 
 
0 #5 Joe Boever 2012-03-13 20:08
I think you need to put some context around your argument that tape is more reliable than disk. If I have a tape sitting offsite and it goes bad, I've now lost data. I also have no idea that the tape is bad until I need the data on it. When a disk goes bad I have no data loss, as that disk is in a RAID set. I also know immediately that it failed because the B2D appliance alerts me.

So yes, on paper an average tape may fail less often than an average SATA disk drive. But practically speaking, in my environment, disk is more reliable for the reasons I've listed above.
 
 
0 #4 W. Curtis Preston 2012-03-12 06:06
Wow, there's a lot of stuff in that one comment.

You removed the questionable statistics, but then linked to an article that uses them. I know of no study that supports your assertions that tape is an inferior medium. I only know of anecdotal-based comments like yours, or articles (like the one you linked to) that say that this person or that person (or thousands of them switched from tape to disk because they didn't like tape).

As to being disingenuous, I wasn't designing to anything, so I don't think so. Having said that, I find the whole thing silly, anyway. We're comparing the prices of two things I don't recommend: using removable 3.5" disks or tape as the initial target for backups.

I will concede that customers do all sorts of things. I will also concede that if you buy a standalone drive from Amazon, they're not going to ask you what you're using it for. And there are probably thousands of people buying standalone tape drives and backing up to them. I'm just saying it hasn't been proper backup design for years. And, yes, I realize that fuels your argument. Remember, despite what you hinted at in your article, I don't have any axe to grind but the truth. Unlike you, I'm not trying to sell anything. (Pretty odd, BTW, to first laud me for honesty and then to suggest in the same article that I've got some vested interest in tape.)

I also don't agree that you and I have done the same thing with pattern recognition. I talked to you about your experiences with tape. What I recall is that all the examples you gave were all very old and do not apply to current technology. My experiences are all with current technology. Disk drives are made to be in a computer or server and were not really designed to be mobile; tapes are designed to be mobile. I am not inferring anything other than these devices work best when they are used for their intended purpose.

Regarding the upgrade issue I'm going to have to say that I am talking about automation environments. I did some googling and found that I could buy an automated tape library for about $2000-2500 that has the capacity of 6-10 TB (native capacity) depending on slots. That's a whole lot more capacity than
we've been talking about for about the same amount of money. Once you go down the automation avenue, you can increase your capacity (assuming that 6-10 TB is enough for a single night's backup) in unlimited quantities for only $.02/GB, which is NOT possible if you're using disks. Your base price is the same as every additional disk you add.

As to the cost of cloud solutions, I think you'll find that they're all over the place. CrashPlanPro is $270/yr for unlimited capacity and 3 computers, or $2,899.00/yr for 1TB & unlimited computers. Carbonite is $1044/yr for 1 TB. Mozypro is over $6,000/yr for 1 TB (Wow). Jungledisk is a service that uses their software and other people's disk (AWS or Rackspace) and is $.15/GB/month plus $5/mth/server. That would be ~$2000/yr for a 5 server 1TB shop. Dollydrive is a Mac-only product that is $288/yr for 1 TB. I looked at bumi.com (an Asigra provider) but I don't see pricing there.

I'm sure you'll look at those prices and say you're cheaper. However, remember to include the daily cost of moving that media around, and storing that media offsite.

Also, all your comparisons to cloud backup seem to think that they all store the backup offsite. Some do, some don't. Some store it in both places. Some store it in three places. Not all cloud backup services are alike. Of the ones discussed above, I know Crashplan, Barracuda, DollyDrive & Asigra support onsite copies. I'm sure there are others. You can actually see Crashplan's GUI showing that here: http://www.crashplan.com/business/features.html#app . Note the Default and Local destinations.

BTW, I'm surprised that you bring up the Google reliability study. It wasn't very complimentary of disk if I recall correctly.
 

Sponsored Links