Google Answers Logo
View Question
 
Q: high-capacity data storage solutions (tape libraries?) ( Answered 5 out of 5 stars,   1 Comment )
Question  
Subject: high-capacity data storage solutions (tape libraries?)
Category: Computers > Hardware
Asked by: joshwa-ga
List Price: $15.00
Posted: 04 Aug 2005 19:01 PDT
Expires: 03 Sep 2005 19:01 PDT
Question ID: 551883
I'm starting up in "bootstrapping" mode (see: seth godin, guy
kawasaki) providing off-site data archiving for my clients, who are in
the media industry. Because I'm dealing with very large volumes of
files (average client: adds 10-15GB/month to a standing archive of
200-400GB), an desktop/server-centric  backup approach is not viable
for me. I'm not backing up PCs, I'm backing up large collections of
files.

requirements:
* (relatively) low entry cost
* expandable up to, say, 20 TB (if I have enough clients to warrant
more than 20TB I'm in a different budget bracket)
* allows for cycling of media off-site (e.g. iron mountain courier
pickup) on a weekly basis (need full off-site media redundancy)
* file-level access
* modest space and cooling requirements, if possible

I'm imagining a tape library, where I can assign each client a tape
magazine or a section thereof, so that way I can swap out whichever
magazine I need to add data to (data added on a monthly or bi-weekly
basis per client).

I'd like to get started on hardware and 1TB (redundant) capacity
without a huge investment (pref <$5K-- is this realistic?)

Is there any reason to implement a SAN over the tape libraries for network access?

What are my hardware/software options for making this happen? Is there
anything I should have considered but haven't?

Clarification of Question by joshwa-ga on 04 Aug 2005 19:03 PDT
forgot one or two things:

* not averse to buying used (ebay, etc)
* LTO vs. DLT? SDLT?
Answer  
Subject: Re: high-capacity data storage solutions (tape libraries?)
Answered By: denco-ga on 04 Aug 2005 20:58 PDT
Rated:5 out of 5 stars
 
Howdy joshwa-ga,

It has been a couple of years since I have delved into serious research
on tape libraries and the such, so I appreciated the chance to update my
personal knowledge on the subject.

The Sony StorStation 152E/A2, as shown on TheNerds.net web site, shows
some promise.
http://www.thenerds.net/index.php?page=productpage&affid=8&pn=LIB152EA2

"The LIB-152E/A2 tape library, with 15-cartridge capacity, holds up to
1.95 TB of data for enterprise level storage requirements."

The above can be easily sourced for under $2,500.

The above uses AIT-2 tapes, which is an 8MM form factor, each capable of
storing 50GB native or 130GB compressed.  Tapes run from $10.00 for
pulls in 100 lots and under $50.00 new.  They have a 64K chip "on board"
that stores the tape's directory, etc. to speed up access.

As you are not opposed to refurbished, etc. units, you could look at one
of the various DLT7000 based units that are out there.

A refurbished Compaq 4120 DLT7000 at Sanworks.com for just under $2300.
http://sanwork.com/product_info.php?products_id=4262

"This high capacity, automated backup solution for NT, NetWare, and SCO
UNIX Servers provide storage in a single unit from 300GB to 1TB."

New IBM 3502-R14 at Compuview Microsystems for under $2800.
http://www.cmishop.com/store/ShowDetails.asp?txtProductName=3502R14-BX

"DLT8000 Tape Automation Products for Netfinity"

A bit over your proposed budget at just over $5700 is the Sony LIB-162/A3,
but it might be worth it.  As shown on the AxionTech web site.
http://www.axiontech.com/prdt.php?src=FG&item=15372

"...
- LEADING STORAGE DENSITY DESIGN PER CUBIC FOOT
- 1.6 TERABYTES NATIVE / 4.16 TERABYTES COMPRESSED
..."

There is no real need to implement a SAN (Storage Area Network) for network
access, at least not at first.  Depending on how complex your future system
becomes, that is something you might have to consider at that time.

This article by W. Curtis Preston on the concept of hybrid tape/ATA/IDE disk
arrays might interest you as well.  From the SearchStorage web site.
http://searchstorage.techtarget.com/originalContent/0,289142,sid5_gci884644,00.html

"Surprise! cheap disks cure slow backup"

I suggest you read the above article in detail, because you could implement
a 1TB server for the interim device between your clients and the Sony tape
library and still be under your $5000 proposed budget.  Fast, flexible,
expandable and relatively inexpensive.

You might also check out this article on the same site.
http://searchstorage.techtarget.com/infoCenter/tip/0,294276,sid5_gci957529_tax294586,00.html

"Guide to low-cost storage"

If you need any clarification, please feel free to ask.


Search strategy:

Google's Froogle search on: terabyte tape library
http://froogle.google.com/froogle?q=terabyte+tape+library

Google's Froogle search on: tape library Sony 152E AIT2 OR "AIT-2"
http://froogle.google.com/froogle?q=tape+library+Sony+152E+AIT2+OR+%22AIT-2%22

Google's Froogle search on: tape library DLT7000
http://froogle.google.com/froogle?q=tape+library+DLT7000

Looking Forward, denco-ga - Google Answers Researcher

Request for Answer Clarification by joshwa-ga on 04 Aug 2005 22:35 PDT
Thanks for your help. A few followups:

* when I want to expand, I just stick another library in the rack and
hook it up via SCSI, right?

* What about Fibre Channel vs. SCSI? iSCSI doesn't really help here, right?

* In order to have on and off-site redundancy, wouldn't I need a tape
duplicator? I need *full* copies of all the data in 2 locations. I
have to cover my own ass while I also cover my clients'. So a
near-line virtual tape library has limited appeal-- speed in restore
is not as high a priority as low TCO.

* Also, am I correct that if I went with a DLT-7000-based system, that
the max I can put in 1 library hovers around 1TB compressed?
(depending on 14 or 15 slots per library)? That means that in order to
achieve my starting point of 2 TB, I have to buy 2 libraries. Which
doubles the price.

* also worth considering is that I am storing media files that don't
compress well, which impacts theoretical vs. actual storage
capacities.

* Where are you finding the bulk pricing on the AIT-2 tapes? I can
only find the single-unit prices which are rather prohibitive. (hey,
there's one case where disks *are* cheaper!)

* The AIT-3 unit indeed looks nice, but is over budget. Perhaps
further down the line. For now I am trying to get the max megabyte for
the buck (can you believe they still measure it that way?)

Eventually I will need to differentiate between on-line, near-line,
and off-line storage, and implement a SAN for performance reasons
(eventually there will be an asset-management front-end for client
access). That's when I'd have to move to a combination of SAN over
RAID, VTL, and tape.

Clarification of Answer by denco-ga on 05 Aug 2005 02:15 PDT
Howdy joshwa-ga,

In the order asked.

> when I want to expand, I just stick another library in the rack and
> hook it up via SCSI, right?

Essentially, yes.

> What about Fibre Channel vs. SCSI? iSCSI doesn't really help here, right?

In a SAN environment you might see some gains using fibre channel versus SCSI.
If you had an all fibre system, transfer rate would be improved remarkably,
but your budget would be shot.

> In order to have on and off-site redundancy, wouldn't I need a tape
> duplicator? I need *full* copies of all the data in 2 locations. I
> have to cover my own ass while I also cover my clients'. So a
> near-line virtual tape library has limited appeal-- speed in restore
> is not as high a priority as low TCO.

This is what could be neat about a hybrid system and a 15 tape system using
one, or better yet, two drives.  Duplicate the information from the server
to two sets of tapes, and do that during "off time" so as to reduce overhead.
A second drive for the Sony LIB-152E/A2 runs around $1000.

> Also, am I correct that if I went with a DLT-7000-based system, that
> the max I can put in 1 library hovers around 1TB compressed?

Yes, which is why, with your budget, the AIT-2 is probably the route to go.

> also worth considering is that I am storing media files that don't
> compress well ...

Again, going with AIT-2 will help with that to a certain extent.

> Where are you finding the bulk pricing on the AIT-2 tapes?

Wholesale Discount Stores has 100 pulls for $10 a piece, or $1000 total.
http://www.wholesalediscountstores.com/product.asp?c=27&s=149&ID=15588

> The AIT-3 unit indeed looks nice, but is over budget.

Ahhh, but so sweet!

> Eventually I will need to differentiate between on-line, near-line,
> and off-line storage, and implement a SAN for performance reasons ...

Agreed.

If you could boost your budget a bit, say to $7000, you could get two
LIB-152E/A2 tape libraries, 100 tapes, and by homebrewing a 1TB server,
you would be sitting in real nice shape in all ways.  5 years ago it
would have cost 10 times as much.  As you know, the real challenge with
your proposal is getting it to work at all for the budget.

We could talk about fibre channel versus Ultra3 SCSIi, etc. forever, but
it always comes back to the bucks.  Even with the suggested setup, and
depending on the number of clients, you are going to have some serious
overtime doing the initial archive for them.

Needless to say, but I will say it anyway, you will want to test, test
and test some more with whatever you end up with, and make sure the backup
and restore process goes smoothly, manageably and efficiently.  Fast would
be nice, but you will need to manage client expectations as well.

Looking Forward, denco-ga - Google Answers Researcher

Request for Answer Clarification by joshwa-ga on 05 Aug 2005 09:21 PDT
> 
> > In order to have on and off-site redundancy, wouldn't I need a tape
> > duplicator? I need *full* copies of all the data in 2 locations. I
> > have to cover my own ass while I also cover my clients'. So a
> > near-line virtual tape library has limited appeal-- speed in restore
> > is not as high a priority as low TCO.
> 
> This is what could be neat about a hybrid system and a 15 tape system using
> one, or better yet, two drives.  Duplicate the information from the server
> to two sets of tapes, and do that during "off time" so as to reduce overhead.
> A second drive for the Sony LIB-152E/A2 runs around $1000.

Two drives is clearly the way to go in my situation.

> If you could boost your budget a bit, say to $7000, you could get two
> LIB-152E/A2 tape libraries, 100 tapes, and by homebrewing a 1TB server,
> you would be sitting in real nice shape in all ways.  5 years ago it
> would have cost 10 times as much.  As you know, the real challenge with
> your proposal is getting it to work at all for the budget.
> 
> We could talk about fibre channel versus Ultra3 SCSIi, etc. forever, but
> it always comes back to the bucks.  Even with the suggested setup, and
> depending on the number of clients, you are going to have some serious
> overtime doing the initial archive for them.

I will be homebrewing a little 400GB portable box to take on site to gather 
the initial data, but as new clients come on-line I'll need to have that in 
reserve.
 
> Needless to say, but I will say it anyway, you will want to test, test
> and test some more with whatever you end up with, and make sure the backup
> and restore process goes smoothly, manageably and efficiently.  Fast would
> be nice, but you will need to manage client expectations as well.

My clients will be asking for single file or subvolume restores, not their 
entire archive at once (and if they did, they'd know not to expect it at 
lightning speed!). 

One last item:

Do you have any software recommendations for managing these libraries, including
the duplicate media sets? I've used Retrospect in the past, but I'm not entirely 
comfortable with its dependence on catalogs (and I shudder to think of the time 
it would take to rebuild the catalog from 40 tapes!)

Thanks for your thoughtful, well-researched answer!
--Joshwa

Clarification of Answer by denco-ga on 05 Aug 2005 13:16 PDT
Howdy Joshwa,

A couple of software packages for you.

LXI - Tape Tracker
http://www.lxicorp.com/prod14.htm

B&L Associates - VaultLedger & Vertices
http://www.bandl.com/products/tape-management-comparison-vaultledger.asp

Dantz - Retrospect - Might have improved since you last used it.
http://www.dantz.com/en/products/win_business/index.dtml

HP - OpenView Storage Data Protector
http://h18006.www1.hp.com/products/storage/software/dataprotector/specifications.html

eMag Solutions - MediaMerge/Tape Management System
http://www.emaglink.com/prodTMS.htm

Looking Forward, denco-ga - Google Answers Researcher
joshwa-ga rated this answer:5 out of 5 stars
Many thanks! Great job!

Comments  
Subject: Re: high-capacity data storage solutions (tape libraries?)
From: denco-ga on 05 Aug 2005 16:59 PDT
 
Thanks for the 5 star rating, Joshwa.

Looking Forward, denco-ga - Google Answers Researcher

Important Disclaimer: Answers and comments provided on Google Answers are general information, and are not intended to substitute for informed professional medical, psychiatric, psychological, tax, legal, investment, accounting, or other professional advice. Google does not endorse, and expressly disclaims liability for any product, manufacturer, distributor, service or service provider mentioned or any opinion expressed in answers or comments. Please read carefully the Google Answers Terms of Service.

If you feel that you have found inappropriate content, please let us know by emailing us at answers-support@google.com with the question ID listed above. Thank you.
Search Google Answers for
Google Answers  


Google Home - Answers FAQ - Terms of Service - Privacy Policy