Veritas-bu

Re: [Veritas-bu] Tapeless backup environments?

2007-09-22 21:06:51
Subject: Re: [Veritas-bu] Tapeless backup environments?
From: "Peter Marelas" <peter_marelas AT symantec DOT com>
To: "Martin, Jonathan" <JMARTI05 AT intersil DOT com>, <veritas-bu AT mailman.eng.auburn DOT edu>
Date: Sat, 22 Sep 2007 17:52:50 -0700
It is interesting to see the points for and against disk / tape backup
technologies play out. A worthwhile discussion.

People have mentioned management/operational/service/infrastructure
costs to justify a switch. Nobody has mentioned risk.

The problem with comparing tape / disk is they are very different
technologies and have different risk profiles subject to how you choose
to apply them.

I don't think disk was every intended to store dormant data. If it did
it would stop spinning don't you think?

So here are some of the risk profile differences to consider before you
take the leap of faith.

1. Tape can be set to dormant / shelved. Disk can not (some can - but
the ones you guys are talking about can't) so it is susceptible to
corruption, malicious intent, accidents, unauthorized and often
undetected access.

2. A tape backup set is isolated from all other tape backup sets - that
is it has few dependents. A disk backup set will often share disks with
others - that is it has many dependents. The risk grows exponentially
with deduping as the logical structure now becomes dependent upon
itself. If I can use an analogy. With deduping your kind of saying
incremental forever to tape is acceptable.

3. It would take a long time to wipe 1000 tapes. It would take a few
minutes to wipe 1000 tape volumes worth of disk and a couple of seconds
for the deduped equivalent.

If deduping was considered risk free we would be deduping our entire
Enterprise. But somehow it is acceptable for backup. I don't think
anyone would agree deduping your backups is acceptable without a tape
backup set. So why do we have deduped backups? Deduping is necessary to
make disk backup viable for a greater share of the backup market.

There is a general rule I apply to technology choices and that is with
every step forward always consider what you are compromising. In this
case it is risk. However, don't get me wrong. The compromise may be
acceptable to you. It is one of those assessments that is difficult to
quantify and therefore often misunderstood or ignored completely.

D2d2T in my mind gives you the best of both worlds (I don't agree one
was every meant to replace the other). It takes away the
unpredictability of tape transports without compromising the data's
resting place and risk profile. The key is in managing the "d" in a way
that affords you tolerance to "T" failures and growth in "D". So to make
it happen the "d" should be far superior to the "D" and "T" combined.

PS. These are my opinions not the companies I work for.

Regards
Peter Marelas
+61 400 882 651

-----Original Message-----
From: veritas-bu-bounces AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Martin,
Jonathan
Sent: Saturday, 22 September 2007 3:37 AM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] Tapeless backup environments?

 
I think what I'm reading here is that no one has done a true 1-to-1
comparison on Tape versus Deduplication / disk.  I guess the next
question is, what would go into such a comparison?

1) Recovery Point Objective
2) Amount of Data To Be Backed Up
3) Retention
4) Cost of Hardware (Deduplication Appliance w/ Disk)
5) Cost of Hardware (Tape Library)
6) Annual Maintenance on Hardware Above
7) Cost of Media w/ Replacement Figures
8) Cost to power / cool disks (infrastructure)
9) Cost of Network link to remote site for de-dupe
10) Cost of Media Transportation and Storage

Price per GB unless factoring in at least all of the above is useless
and much of that information depends on configuration.  I did such an
analysis when we upgraded to NBU6 and considered deduplication this time
last year.  In my case, many of the features of disk based deduplication
weren't applicable to my situation (especially RPO) so tape was easily
cheaper.  If you are shipping media offsite daily though for a >=1 day
RPO then deduplication definitely makes a play.  Further price per gig
on the disk side has been heavily influenced by "consumer grade" SATA
drives at 750gb and 1TB bringing costs way down in comparison to only 1
or 2 years ago.

There's certainly a lot of data to injest before making claims of either
technology's superiority in a particular environment.

-Jonathan


-----Original Message-----
From: veritas-bu-bounces AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Curtis
Preston
Sent: Friday, September 21, 2007 1:10 PM
To: Justin Piszcz; Jeff Lightner
Cc: veritas-bu AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] Tapeless backup environments?

First, you can't compare the cost of disk and tape directly like that.
You have to include the drives and robots.  A drive by itself is useful;
a tape by itself is not.  

Setting that aside, if I put that disk in a system that's doing 20:1
de-duplication, my cost is now 1.65c/GB vs your 3-9c/GB.  

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

-----Original Message-----
From: veritas-bu-bounces AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Justin
Piszcz
Sent: Friday, September 21, 2007 7:36 AM
To: Jeff Lightner
Cc: veritas-bu AT mailman.eng.auburn DOT edu
Subject: Re: [Veritas-bu] Tapeless backup environments?


I believe disks are 33c/gigabyte and tapes are 3-9cents/gigabyte or even

cheaper, I do not remember the exact figures, but someone I know has
done a cost analysis and tapes were by far cheaper.  Also something that
nobody calculates is the cost of power to keep disks spinning.

Justin.

On Fri, 21 Sep 2007, Jeff Lightner wrote:

> Disk is not cheaper?  You've done a cost analysis?
>
> Not saying you're wrong and I haven't done an analysis but I'd be 
> surprised if disks didn't actually work out to be cheaper over time:
>
> 1) Tapes age/break - We buy on average several hundred tapes a year - 
> support on a disk array for failing disks may or may not be more 
> expensive.
>
> 2) Transport/storage - We have to pay for offsite storage and transfer
-
> it seems just putting an array in offsite facility would eliminate the

> need for transportation (in trucks) cost.  Of course there would be
cost
> in the data transfer disk to disk but since everyone seems to have 
> connectivity over the internet it might be possible to do this using a

> B2B link rather than via dedicated circuits.
>
> 3) Labor cost in dealing with mechanical failures of robots.   This
one
> is hidden in salary but every time I have to work on a robot it means
I
> can't be working on something else.   While disk drives fail it
doesn't
> seem to happen nearly as often as having to fish a tape out of a drive

> or the tape drive itself having failed.
>
>
> -----Original Message-----
> From: Justin Piszcz [mailto:jpiszcz AT lucidpixels DOT com]
> Sent: Friday, September 21, 2007 10:08 AM
> To: Jeff Lightner
> Cc: veritas-bu AT mailman.eng.auburn DOT edu
> Subject: Re: [Veritas-bu] Tapeless backup environments?
>
>
>
> On Fri, 21 Sep 2007, Jeff Lightner wrote:
>
>> Yesterday our director said that he doesn't intend to ever upgrade 
>> existing STK L700 because eventually we'll go tapeless as that is
what
>> the industry is doing.   The idea being we'd have our disk backup
>> devices here (e.g. Data Domain) and transfer to offsite storage to 
>> another disk device so as to eliminate the need for ever transporting

>> tapes.
>>
>> It made me wonder if anyone was actually doing the above already or
> was
>> planning to do so?
>>
>
> That seems to be the way people are 'thinking' but the bottom line is 
> disk still is not cheaper than LTO-3 tape and there are a lot of 
> advantages to tape; however, convicing management of this is an uphill

> battle.
>
> Justin.
> ----------------------------------
>
> CONFIDENTIALITY NOTICE: This e-mail may contain privileged or
confidential information and is for the sole use of the intended
recipient(s). If you are not the intended recipient, any disclosure,
copying, distribution, or use of the contents of this information is
prohibited and may be unlawful. If you have received this electronic
transmission in error, please reply immediately to the sender that you
have received the message in error, and delete it. Thank you.
>
> ----------------------------------
>
>
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu