Comments on: Where is Data Deduplication going? http://thesantechnologist.com/?p=37 A Storage Technology Blog by a Technologist Mon, 26 Oct 2009 04:20:58 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: Follow-Up to Deduplication… | The SAN Technologist http://thesantechnologist.com/?p=37&cpage=1#comment-299 Follow-Up to Deduplication… | The SAN Technologist Sat, 27 Sep 2008 19:12:31 +0000 http://thesantechnologist.com/?p=37#comment-299 [...] I’ve written a couple times about Deduplication, mainly in regards to my feeling that it is a feature and not a product, and more recently looking at NetApp’s implementation of A-SIS.  I also mentioned the [...] [...] I’ve written a couple times about Deduplication, mainly in regards to my feeling that it is a feature and not a product, and more recently looking at NetApp’s implementation of A-SIS.  I also mentioned the [...]

]]> By: Steven J. Schwartz http://thesantechnologist.com/?p=37&cpage=1#comment-246 Steven J. Schwartz Thu, 04 Sep 2008 18:03:16 +0000 http://thesantechnologist.com/?p=37#comment-246 David, Sorry this has taken so long to respond to, your comment got caught in my comment SPAM filter. The issue with ASIS from NetApp is varied. It actually is a POST Production process, and it has huge performance impacts to the volumes being de-duplicated. It can also ONLY dedup at the VOLUME level, not at the filer level, so depending on your configuration you might not get the expected return. David,

Sorry this has taken so long to respond to, your comment got caught in my comment SPAM filter. The issue with ASIS from NetApp is varied. It actually is a POST Production process, and it has huge performance impacts to the volumes being de-duplicated. It can also ONLY dedup at the VOLUME level, not at the filer level, so depending on your configuration you might not get the expected return.

]]> By: David Henderson http://thesantechnologist.com/?p=37&cpage=1#comment-82 David Henderson Fri, 09 May 2008 20:04:00 +0000 http://thesantechnologist.com/?p=37#comment-82 This is a relevant discussion for me since right now I am at a fork in the road. We have two NetpApp FAS3020 filers. One of my filers is for Primary storage and I mirror to the second one which is in my disaster recovery site. We use Commvault to backup the data on my DR filer to tape as well as data from the C: drive on all of my servers. I am looking at replacing tape with disk based backup and have been looking at two options: - Getting another disk shelf for my NetApp and uses SnapVault to do the backup and ASIS for dedup - Purchase and Exagrid for backup Any thoughts? This is a relevant discussion for me since right now I am at a fork in the road. We have two NetpApp FAS3020 filers. One of my filers is for Primary storage and I mirror to the second one which is in my disaster recovery site. We use Commvault to backup the data on my DR filer to tape as well as data from the C: drive on all of my servers.

I am looking at replacing tape with disk based backup and have been looking at two options:
- Getting another disk shelf for my NetApp and uses SnapVault to do the backup and ASIS for dedup
- Purchase and Exagrid for backup

Any thoughts?

]]> By: Adam http://thesantechnologist.com/?p=37&cpage=1#comment-79 Adam Wed, 30 Apr 2008 02:57:43 +0000 http://thesantechnologist.com/?p=37#comment-79 I appreciate your opinion, although, I strongly disagree. I was at VMware when people were saying that it was just a fad and I don't think that the internet will go away any time soon either. Your market share numbers are not accurate. Data Domain's competitors won't be the big companies like Netapp/EMC as the big companies have their feet set in the cement 10-30 years ago with product cores such as Flare, WAFL/Ontap, Nearstore. The big OEMs are handcuffed by these product cores that are very complex, old and inflexible making it impossible to do much more than spend lots of money marketing a "pig with lipstick" and give it away for "free" (this didn't work for Microsoft with VMware and Citrix didn't go away when they announced free TS). Many opinions suggested Netapp was going to die in near Y2K because EMC launched the IP4700 but we all know that nearly a decade later EMC is still trailing the capabilities of what Netapp is able to do with respect to file serving, mgmt, deployment, snap capabilities. It is truly not as simple as you are suggesting to develop relevant products that meet customers requirements - and customers do talk. Here is what I think will happen in the space. Big guys will come out with weak products developed internally or will go out and buy some of the startups that have a product that doesn't work good enough to sell to customers as fast as they are burning through their VC cash. Now with the credit crunch, VC funding has dried up and most of the startups need to sell more products to fund investment/growth or sell the company as they can't make it on their own. Most of the small startups are now in the "sell" position - Diligent was just picked up last week for $160M-200M by IBM and their are rumours about Exagrid having trouble getting VC cash. Falconstore in being squeezed out of all of the OEM's plans, Sepaton will likely be picked up by HP and Dell may grab Exagrid - some rumours are floating around currently. Noboby will touch Quantum as they have more long-term debt than the value of their market cap (other than EMC sucking their blood like a mosquito on a dead carcus through their up and coming OEM annoucement). Netapp has a ton of arrogance internally and will keep coming out with home-grown diluted products that will disappoint their customer base and the same thing will happen to them that happened to EMC in the space - their core business will soften as companies like DD, Equallogic, 3Par, Compellent, Copan build better mousetraps. EMC was smart when this began to happen as Tucci decided to start a massive product portfolio diversification by acquiring about a billion companies. Back to Data Domain. Clearly at 1.3B market cap, not a lot of companies can afford to buy the company other than maybe a CISCO (lots of interesting reasons to do this but still not likely). The deployment of the technology is in-line, always de-duping, relevant to backup/recovery/DR and archiving and they have found a way to make it fast. Customers love it and they are adding new customers faster than any other company is in this space by an order of magnitude. They will continue to eat away at the backup/archive/tier 2 disk market as Netapp did to EMC through the 90's and 00's. My last comment is that the best technology does not necessarily always win with customers so Data Domain's job will be to persistently execute in the market, educate prospective customers and convince their customers to spread the word. My prediction is that Data Domain will become the Netapp of the new millenium. I appreciate your opinion, although, I strongly disagree. I was at VMware when people were saying that it was just a fad and I don’t think that the internet will go away any time soon either. Your market share numbers are not accurate. Data Domain’s competitors won’t be the big companies like Netapp/EMC as the big companies have their feet set in the cement 10-30 years ago with product cores such as Flare, WAFL/Ontap, Nearstore. The big OEMs are handcuffed by these product cores that are very complex, old and inflexible making it impossible to do much more than spend lots of money marketing a “pig with lipstick” and give it away for “free” (this didn’t work for Microsoft with VMware and Citrix didn’t go away when they announced free TS). Many opinions suggested Netapp was going to die in near Y2K because EMC launched the IP4700 but we all know that nearly a decade later EMC is still trailing the capabilities of what Netapp is able to do with respect to file serving, mgmt, deployment, snap capabilities. It is truly not as simple as you are suggesting to develop relevant products that meet customers requirements – and customers do talk.

Here is what I think will happen in the space. Big guys will come out with weak products developed internally or will go out and buy some of the startups that have a product that doesn’t work good enough to sell to customers as fast as they are burning through their VC cash. Now with the credit crunch, VC funding has dried up and most of the startups need to sell more products to fund investment/growth or sell the company as they can’t make it on their own. Most of the small startups are now in the “sell” position – Diligent was just picked up last week for $160M-200M by IBM and their are rumours about Exagrid having trouble getting VC cash. Falconstore in being squeezed out of all of the OEM’s plans, Sepaton will likely be picked up by HP and Dell may grab Exagrid – some rumours are floating around currently. Noboby will touch Quantum as they have more long-term debt than the value of their market cap (other than EMC sucking their blood like a mosquito on a dead carcus through their up and coming OEM annoucement). Netapp has a ton of arrogance internally and will keep coming out with home-grown diluted products that will disappoint their customer base and the same thing will happen to them that happened to EMC in the space – their core business will soften as companies like DD, Equallogic, 3Par, Compellent, Copan build better mousetraps. EMC was smart when this began to happen as Tucci decided to start a massive product portfolio diversification by acquiring about a billion companies.

Back to Data Domain. Clearly at 1.3B market cap, not a lot of companies can afford to buy the company other than maybe a CISCO (lots of interesting reasons to do this but still not likely). The deployment of the technology is in-line, always de-duping, relevant to backup/recovery/DR and archiving and they have found a way to make it fast. Customers love it and they are adding new customers faster than any other company is in this space by an order of magnitude. They will continue to eat away at the backup/archive/tier 2 disk market as Netapp did to EMC through the 90′s and 00′s.

My last comment is that the best technology does not necessarily always win with customers so Data Domain’s job will be to persistently execute in the market, educate prospective customers and convince their customers to spread the word. My prediction is that Data Domain will become the Netapp of the new millenium.

]]>