Veritas-bu

Re: [Veritas-bu] Data Domain vs Quantum DXi

2008-11-09 04:05:26
Subject: Re: [Veritas-bu] Data Domain vs Quantum DXi
From: "Curtis Preston" <cpreston AT glasshouse DOT com>
To: "Stafford, Geoff" <GStafford AT barclaycardus DOT com>, <Robin.Small AT fresno DOT gov>
Date: Sun, 9 Nov 2008 00:59:28 -0500
>>NAS/CIFS sucks for
>>any backup you actually care about throughput on IMHO.

OST isn't NAS/CIFS.  It's a completely different protocol.  Tests on DD
show it to be at least 100% faster than CIFS.  (FWIW)  Also, it can be
done over Fibre Channel, as some other vendors are working on.

>--It's a variable length block size deduplication algorithm, the math
is
>above my pay grade but there's probably only so many ways you can do it
>efficiently and apparently Riverbed thought it was bs and wanted it to
>go away too.  :)  No one from any company has ever told me there was
>shared code, I was simply working on the KISS principle in my previous
>email and not trying to spend 10 paragraphs explaining it hence the
>"short version of the story" comment.  :)

But the short version you told isn't a summary of what happened.  They
aren't licensing anything from Quantum.  Their method was close enough
that Quantum thought they could sue them on patent infringement, and DD
settled to put an end to the lawsuit.  (As did Riverbed.)  None of that
translates into "DD is licensing Quantum's technology," which was your
short version.  Neither, IMHO, should the settled lawsuit issue
translate into the deduction you made that Quantum would enhance the
technology before DD would.

I objected to the "short version" so strongly because Quantum sales reps
apparently continue to propagate it as a sales tactic, and that's just
BS.  IMHO, it shouldn't in any way factor into your decision on which
product to buy.

>>Remember that just because you have the capacity doesn't mean you can
>>use it.  DD stops their boxes at certain capacities because most
>>customers run out of bandwidth before they run out of space.

>>>--Depends on how you want to use it and how long you plan on
retaining
>>>data on disk.  YMMV

While I agree that YMMV (especially in dedupe), I don't agree that this
is one of those areas.  The problem is that if you store that much stuff
behind a single head, the performance can get so slow that the head
becomes unusable.  (The bigger the datastore, the bigger the hash table.
The bigger the hash table, the longer hash lookups take, the worse
performance gets.)  NOW... If you can fit 200 TB (or whatever number you
gave) in that thing and have the performance stay acceptable, I'll
change my mind.

>> Agreed, but that functionality is supposed to ship RSN.  And Quantum,
>>AFAICT, hasn't shipped it either.  It's in the code but hasn't
shipped.

>>--At what cost for the functionality on the Data Domain side?  

I'm not sure they know yet.







This email and any files transmitted with it are confidential and intended 
solely for the use of the individual or entity to whom they are addressed. If 
you have received this email in error please notify the system manager. This 
message contains confidential information and is intended only for the 
individual named. If you are not the named addressee you should not 
disseminate, distribute or copy this e-mail.

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