VTL vs. Backup to Disk

"Joe Rom King" writes:

!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"

Can you please turn off HTML posting? Some of us don't have news
readers that can handle it.

--
David Magda dmagda at ee.ryerson.ca
Because the innovator has for enemies all those who have done well under
the old conditions, and lukewarm defenders in those who may do well
under the new. -- Niccolo Machiavelli, _The Prince_, Chapter VI

The bigggest benefit I know of is the ability to quickly get your data
backed up . If you need to store offsite copies of tape the best solution I
found is the ADIC Pathlight VX. It automates the path to tape making it
effortless. I backup to disk then let the VX handle when teh data gets old
the migration to tape. At the same time the data is first acked up to disk I
also have the ability to let the ADIC device cut a copy of tape for offsite.
So in essense I have 3 copies of my data. The big difference between this
and aother D-D-T is I only need one library license from Veritas unlike the
CDL option from EMC where you would need two library licenses.

wrote in message
oups.com...
What other considerations might there be. The two environments I'm
speaking of would have their own dedicated disk/VTL and sharing
wouldn't be a issue. What about cost, performance or functionality?

These are all good points.

Disk is designated as DASD (direct access storage device). This means
that the disk can do random reads and random writes. Tape can only do
random reads without truncating the data. This functionality must be
emulated to provide a complete tape emulation solution.

Both are block devices. It does not take much emulation to make the disk
look like a tape; in fact there is a patent that has been around for
many years covering tape emulation using disks.

The question boils down to backup application support. Most of the major
backup ISVs support tape and tape libraries. Disk-to-disk support is
sketchy, so a good VTL may provide for a better solutions and better
compatibility with existing backup applications.

The idea of a changer is not restricted to tape. There are numerous CD
and DVD changers out that work similar to the way tape changers work; in
fact the SCSI specification for a changer is not bound to tape.

I suspect for an ISV to do D2D properly, a changer emulation layer needs
to be implemented in the D2D environment. For now, I recommend using a
VTL. NovaStor also has a VTL package. I have not used it, so I can not
comment on it.

Jamey Kirby


-----Original Message-----
From: ]
Posted At: Friday, August 12, 2005 7:14 AM
Posted To: comp.arch.storage
Conversation: VTL vs. Backup to Disk
Subject: VTL vs. Backup to Disk

Brian,
Here is some stuff I did when I was an analyst w/ ESG. It talks a
little bit about the difference.

Also, it may add a little on the cost, but on the complexity side,
after setup (configuring the thing to look like the tape library you
want) it is very simple to deploy, manage and recover.

The cost may be up there w/ things from EMC or NTAP but there are some
good prodcuts available from a number of different vendors that are
less.

here are some of my writings

Does Disk Solve The Problem?
Based on our last comments about integrated disk and tape solutions,
and all the hype disk based backup solutions are getting these days,
lets take a closer look at how disk based solutions are helping the
data protection paradigm evolve. There has been a lot of talk
concerning how the backup paradigm will change toward disk-based
solutions. The latest and most common discussion is about how
disk-based technologies will be used to increase data recovery. As ATA
arrays are becoming more mainstream and commoditized, the cost per MB
of low-end disk arrays are beginning to converge with tape costs. As
this convergence happens, IT managers are starting to discuss the
cost-related pros and cons of managing their existing backup processes
versus the ease of data management and higher availability that
disk-based backup solutions provide. Again, most of these discussions
revolve around disk and disk-based appliances.

Traditional Backup/Recovery Technologies - Most of the traditional
software backup vendors (BakBone, CA, CommVault, HP, Legato, SyncSort,
Tivoli, and Veritas) have stated that they do or will support backup to
disk. This means that IT can now purchase one of the many new ATA
arrays to be used for caching their tape libraries. Moving data to
disk with a traditional backup product is still backup in its purest
form. The faster speed of disk has helped with shrinking backup
windows, but only slightly.

To the credit of the traditional backup vendors is that they are
continuing to innovate. Companies such as Veritas are adding the
capability to do synthetic full backups. This helps alleviate the need
for full backups. BakBone continues to innovate as it pertains to
networked storage backup, also a good platform for increasing backup
performance.

Data growth is still outpacing this backup window, which leads us back
to the theme of this brief, which is, "How will data protection
evolve to meet these challenges?"

Virtual Tape Libraries (VTL) - Over the past 18 months, we have seen
more and more intelligence move from the backup software into either
the fabric or the arrays being used for backup/recovery. Recently,
there has been a spate of VTL announcements. VTL software is designed
to make a disk array look like a tape library. Vendors who deliver such
technology include Alacritus, Diligent, Falconstor, IBM, Quantum,
Sepaton, and even BakBone (which has designed VTL capabilities into its
software). Some of these companies deliver software while others
deliver appliances with software built in, such as the Quantum DX30.
The goal with VTL is to get IT past the psychological inertia
associated with migrating from disk- to tape-based backup. VTLs fit
seamlessly into the existing backup environment and require very little
or no change to the existing processes. The only change in the backup
process is for IT to redirect the backup application to the IP address
of the array instead of the tape library, and backups should be ready
to run. The data written to disk is exactly what you would expect to
see on tape, and can be easily cloned or moved quickly to tape without
affecting media management schemas - which is crucial when it comes to
recovery.

Disk Backup/Recovery Targets - As we move along the data protection
continuum, there are a couple of diverging schools of thought on
disk-based backup and recovery. In the first camp is the group of
people who believe in disk-based backup as cache and don't understand
the need to put the data on disk in a format that is like tape. Their
view is, "What is the point? VTL software costs more money and at
the end of the day, IT must have a process of moving the same volume of
data off to tape for archival." However, depending on the backup
software, cloning data to tape may not be possible when VTL is not
used. For example, data written to a 120GB disk drive with backup
software cannot be cloned to 60GB tapes. Trying it could also cause an
issue with the media management schemas.

This is where the conflict between existing processes and new processes
comes into play. While some IT folks attempt to mirror existing
processes, others only care that their primary storage is always
available, and they don't care that it may take all day to move backups
from the ATA array to tape. That said, most disk vendors' products
can act as a target for backup, these vendors include Iomega, Network
Appliance, NexSAN, and Snap as well as many others.

The other disk-based school of thought deals with recovery. This is
not to say that by putting data on disk, you don't increase recovery
speed, it does, but IT still needs to use the processes defined by
traditional backup software to recover data. Even though recovery
focused technologies may not meet all backup requirements, they can
help companies move up the data protection index by promising faster
backup and more importantly, a greater chance of data recovery.