Networker

Re: [Networker] VTL or disk cabinet backup

2007-11-04 23:28:22
Subject: Re: [Networker] VTL or disk cabinet backup
From: Curtis Preston <cpreston AT GLASSHOUSE DOT COM>
To: NETWORKER AT LISTSERV.TEMPLE DOT EDU
Date: Sun, 4 Nov 2007 23:23:51 -0500
Oscar makes some valid points about his concerns about VTLs.  Some of
them apply to all VTLs, some of them apply only to one or a few of them,
and at least one doesn't apply to any VTLs of which I'm aware.  The VTL
industry has brought this FUD (fear uncertainty & doubt) upon
themselves, as they often resort to FUD throwing to make the sale.
Hopefully I can clear that up.  (I don't work for any of these vendors,
but I work WITH all of them.)

As I'll say later in this email, I'm not sure VTLs DO add very much in
small environments, or that they have a major advantage in backup
environments that don't have multiple servers, but I do believe they add
value in very large environments.  I'll explain that later.

First let me say that for any environment backing up multiple terabytes,
I believe that using a non-dedupe disk device as your primary storage
for backups is at this point a waste of money.  Dedupe brings too much
value to not use it in a disk target that's going to be used for
anything other than staging.  (Note: I'm not talking about disk staging
here.  I'm talking about using disk as your primary onsite storage
device.)

I'm assuming, then, you're going to be buying a dedupe device.

>* VTLs can support data deduplication
> - And so can raw disk with additional hardware.

Today (and for the foreseeable future) there are only two types of
dedupe devices: VTL and NAS heads.  That means that if you want a block
device that you can communicate to without using IP, then you're going
to want VTL.  If you're OK with IP/NAS-level performance, then either
NAS or VTL can satisfy your performance requirements.  Current data
suggests that a dedupe VTL and dedupe NAS head add about the same cost
to the disk, so then the question is: which brings more value?

>However, data dedup loses 
>its appeal as it can't handle several save streams to one device at
once.

I think you're saying that dedupe devices can't support multiplexed
backups and maintain their dedupe ratio.  This is NOT true of all dedupe
devices; it is true of some of them.  (Ask your vendor.)  In addition,
I'd ask "why are you multiplexing to a disk device?"  There's no reason
to do that in a NetWorker environment.  If you want to do 40 jobs
simultaneously, create 40 virtual tape drives, not 10 virtual tape
drives with 4 jobs each.

>And why would I want to put an expensive and at the same time limiting 
>emulation layer between disks and software, as long as my software can 
>support disk-based backups?

The biggest reason is de-dupe.  (Other reasons come later in this
email.)  And the only way to get that today is to buy a NAS head with
dedupe or a VTL head with dedupe, and the data I have shows that they're
about the same price.  If they're not, everything's negotiable.  

So then the question is: which works better for you?  NAS or VTL?  This
one gets a little more difficult to discuss in this forum as it needs
lots of drawings and such, but suffice it to say that I believe that the
larger your environment is, the more you're going to want to LAN-free
backups, and LAN-free backups can't be done to a NAS device (it's on the
LAN).

>* VTLs can help you cirrcumvent networker licensing for AFTDs.
> - Considering both that VTLs impose severe limitations on what can be 
>done with disk, at an additional cost to that, this argument becomes 
>invalid. Besides, isn't EMC going to create specific VTL licenses too?

While I don't agree that VTLs impose severe limitations, I do agree that
this isn't a good reason to choose VTLs.  The only limitation I'm aware
of is the lack of simultaneous read/write, but I agree with another
poster that said that this can be handled with proper configuration.
(Basically smaller virtual tapes, but that does frustrate the licensing
issue.  I hope that NW comes out with VTL-friendly licensing soon.)

>* VTLs can clone virtual tapes to physical ones, thus ofloading I/O
from 
>the server.
> - Buy a better server then? Besides, I assume that cloning of virtual 
>tapes to physical ones put a constratint on how much data can be
squeezed 
>into physical media, as I assume its hard to calculate how much 
>compression can be done when the data is written to physical tape.

I would totally agree with you here.  FWIW, other backup products are
working with some VTLs & NAS products to figure out how to do this
without the drawbacks you mentioned.  (Already GA in NetBackup.)

>So, the conclusion is: Don't bother with VTL. Get a cheap JBOD, and use

>ZFS, or a cheap RAID-capable box (InforTrend EonStor, Fujitsu-Siemens 
>Fibrecat SX88 or similar) and invest your money in a diskbackup license

>instead.

If all you want to do is disk staging, I agree with you for certain size
environments.  However, large environments are going to find a lot of
value in dedupe and multi-head VTLs.

>One other thing that's worth noting about diskbackup licenses for 
>networker is that the license puts restrictions on the actual amount of

>data on disk, compared to library based licenses, who put restrictions
on 
>the amount of slots instead. This means that unlike library licenses, 
>diskbackup licenses don't allow you to put more data in over time, as 
>technology evolves.

That's an interesting thought.  It's a drawback of capacity-based
licensing.

What value do _I_ think that VTLs bring to the table?  In a small
environment, I'm not sure they add much.  (I don't think they subtract
much either.)  In medium-sized environments (one NW server, NO storage
nodes), the addition of dedupe really gives you a reason to move off
JBOD/RAID and onto an intelligent disk device.  In such environments,
most of the midrange dedupe NAS or VTL products will work.  In large
environments (multiple NW servers, storage nodes, perhaps other backup
products as well), I'd say that VTLs win hands down over either standard
disk or dedupe NAS.

Standard disk is MUCH harder to use in large, multi-backup-server
environments because of all the provisioning issues.  You have to create
and manage one or more RAID volumes per backup server/storage node, etc.
You're always going to have volumes that are too big or too small,
creating something else to manage.  You can't easily move backups from
one backup server/storage node to another.  If you have different Oss,
you can't even mount the RAID group/volume you made on OS to another
OS's backup server.  You get no dedupe or hardware compression.  You
will have fragmentation issues if you use them as a permanent storage
device (as opposed to disk staging.)

NAS disk (dedupe or otherwise) doesn't meet the needs of very large
environments either, as many of the servers that need to be backed up
need LAN-free backups.  LAN-free backups mean using a block device, and
today a block device means either standard disk or dedupe VTL.  I've
already said what I thought about standard disk, so that leaves only
VTL.

VTLs can easily be shared between multiple backup servers, storage nodes
-- even applications that don't share -- without creating and managing
individual RAID volumes for each server.  They have dedupe and hardware
compression, and any good dedupe device has worked out the fragmentation
issue as well.

---
W. Curtis Preston
Backup Blog @ www.backupcentral.com
VP Data Protection, GlassHouse Technologies

To sign off this list, send email to listserv AT listserv.temple DOT edu and 
type "signoff networker" in the body of the email. Please write to 
networker-request AT listserv.temple DOT edu if you have any problems with this 
list. You can access the archives at 
http://listserv.temple.edu/archives/networker.html or
via RSS at http://listserv.temple.edu/cgi-bin/wa?RSS&L=NETWORKER