Does NetApp’s purchase of LSI’s storage business – not known as Engenio – portend a fundamental change in NetApp’s business model? No and yes. Here’s why.
The story
NetApp has often contrasted their single focus on Data ONTAP with EMC’s confederation of businesses. NetApp may be much smaller than EMC, but their focus on a single software stack has given them an R&D and marketing advantage over EMC.
If that is changing, customers are right to be concerned. If NetApp is backing away from DOT after a painful and lengthy rewrite, that would suggest that all is not well.
No
The LSI buy does not represent a major strategic shift. Why? Because there is little overlap between the NetApp and LSI product sets. The new product line is additive, not cannibalization.
The LSI product line is focused on fast direct-attached storage. The controllers are popular in video and HPC. NetApp, of course, is big on SAN and NAS.
Yes
While it isn’t a major strategic shift, the LSI buy is a shift: they are adding to their product portfolio to win incremental customer dollars. They are also adding an OEM business to their already healthy reseller channel.
The StorageMojo take
New CEO Tom Georgens continues to push overdue changes. The LSI buy is an easy win for NetApp: they already have a storage sales force and a wide enterprise footprint.
Adding fast DAS to the portfolio is smart: the sales and support teams can easily train-up on the LSI kit without losing focus on their bread and butter. With direct sales NetApp can apply proven end-user Margin Enhancement Technology to Engenio’s anemic OEM margins.
The more interesting test is seeing how Georgens integrates Bycast into the product portfolio. Will it be positioned to serve a new market, or will it be blended with existing products?
The bottom line is that NetApp hasn’t had many – any? – successful acquisitions. It is a skill they need to learn.
They needn’t imitate EMC’s technology publishing model – of which Documentum and RSA are unneeded distractions – using their financial clout to purchase brands that have crossed the chasm. EMC’s world-wide sales coverage allows them to pay top dollar and then rapidly pump up sales.
NetApp isn’t in that league – yet. Baby steps now will pay off in the long run.
Courteous comments welcome, of course. I was impressed Tuesday night by NetApp employee support for the St. Baldrick’s Foundation benefit for pediatric cancer research.
I sent a note to a local engineer I know in my area that works for Bluearc, they are pretty big users of that LSI stuff, heavily promote it for customers that can’t afford the HDS disks for whatever reason. I can’t imagine what might be going through their heads, the prospect of a big competitor of yours now being one of your suppliers, can’t be a good feeling!
It may be a boon for others like Infortrend which seem to offer similar storage technology to LSI (at least on the surface).
Also makes me wonder about the future of Onstor which LSI bought, a friend of mine in the industry was convinced LSI was going to integrate Onstor NAS into their Engenio systems and have a nice cost effective multiprotocol system but apparently that never materialized. And Onstor hasn’t been shown much love in other areas of LSI, while the documentation seems to be behind registration walls now last time I looked(few months ago) it was pretty shocking to me, almost across the board the Onstor product did not support online back end storage upgrades, from pretty much any storage system (talking about SAN upgrades here not the Onstor upgrades itself), whether it was IBM storage, HDS, 3PAR and even many of LSI’s own storage systems the official recommendation was to disconnect the Onstor when doing software upgrades on the SAN.
Not that Onstor was that hot of a product to begin with but I was kinda-sorta expecting LSI to put a bit more resources into the thing. At one point in a casual quoting exercise I went through with Onstor last year they were apparently down to only a handful of engineers and a single sales guy for the whole thing.
Talking of big strategy shifts, with today’s Oracle announcement they’re ditching Itanium, i feel we’re on a cycle back to vertical single supplier stacks.
If i didn’t have skin in my company, i’d probably be a certifiable idiot, for mucking around in the trenches. Corollary, spent a good decade or two doing just that to get the skin in the game. Either way the whilst i’m hoping the proverbial option pays my pension, the option is putting in for seniority promotion 🙂
Some months i check in here and to the couple other notable blogs and go “Aha! That’s what’s up.” This month, being pressed with fairly serious LOB things, my eyes are glazing over.
What i think i see is hyper differentiation. Everyone does “this but not that and kinda that but no, you’re not our target vertical”. How do you pitch that? It’s unadulterated cacophony. Within companies now, because of acquisitions.
But these new companies are going to be mincemeat if the verticalization trend continues. I’m including NetApp here, as they’re most likely to have the scale and attitude to slug it out.
OK, i’ll accentuate the positive and say the hypotheticals (of spending boatloads on new storage products) leads back to some very detailed new thought at our application level. We’re lucky, the LOB transactional stuff was always kept simple and well divorced from analytics e.g. Coming to that in a mo. as relevant.
Philip Greenspun has been blogging about relative salaries, asking hypothetically how male programmers get spouses, comparing salary surveys to other highly competent / skilled professionals. Worth a look. This is relevant, because there’s this whole mismatch where CIO says to board, “yeah but it’s complex” so hires a consultant, consultant outsources, *everyone outsources* and the pay grades are held down by effective conspiracy. The sheer inefficiency is mindboggling, but i’m preaching to the choir. My own niche business area suffers from that with bells on.
But i drift. To try to tie these disparate thoughts together.
– So someone pitches me storage which will run my DB / app n* faster.
– Months and infinite blurb and waffle later, I walk off and look at the code, and start getting the gain for time and typing. (remember the code divorce, an early lucky “decision”)
– I get frustrated enough, i do get tempted to go cap in hand and beg to be molested by one company. I might even do so, if I believed a SLA’s legalese.
– The alternative is to bring in a small bright consultant but this is more time, plus the time to legally iron ball and chain them on sight of design specs, before they start work. By that time a straight hire would make more sense for all.
– My medium term plan is to work out some long dated option plan that can be reasonably expected by a normal person and not a quant, to golden handcuff new staff, even if they leave, and bring them in line with a modern real salary along the way.
– Overriding rationale. I could neither rework the code nor understand requirements if i couldn’t shout at the authors 🙂
I’m a big fan of DAS too. There is too little choice there, and far too much cloud “but we support http and direct URI object paths (if you supply a nice object layer yourself)!” rubbish talk.
Now, will NetApp take the suggestions i’m mooting above and hoping to follow, so they have a chance of actually integrating what they’ve bought? It’s hardly new thinking. If they do, i’ll make the call. Meanwhile, my news comes from here 🙂
This jungle of cloud stuff (being dismissive for effect, not handwaving the subtleties) is making mockery of the pay grades of who have to genuinely support it. They are being lost in between corporate PR wars and CxO chest bumping, and despite apparently good rates, don’t – from where i’m looking – get the opportunity to make really good money in a long term position.
(and that’s the good guys. I live London, where you can live like a hermit, if you’ve no family, you can take a hellish commute – and i mean hellish as the next decade will see the entire transit infrastructure dug up and repaired – and where my 1 bed central apartment i long ago sold, recently traded for just shy of USD$ 2 million.)
My apologies for the rant factor 9 (log) above. I just want to shout loud that i think this move by NetApp is pretty good and that what’s lacking is cutting through the fetid undergrowth of the world they’re selling to. Good strategy involves the people outside your organization.
all best to all,
– john
For clarity:
“I’m including NetApp here, as they’re most likely to have the scale and attitude to slug it out.”
please mentally add, “And loose. Because they can be cut out, just as Itanium with Oracle.”
I think that by the acquisition of LSI Engenio product line, NetApp is maybe trying to stop another trend which is called OpenStorage and has been started by Nexenta with their NexentaSTOR product, as you may read here: http://www.nexenta.com/corp/newsflashes/102-2011/900-lsi-buy-may-disrupt-openstorage-channel-growth
Hi Alessandro,
Do you mean, that by taking out one kind of building brick, that NetApp might choke growth in whitebox kit?
I don’t really see that happening, if that’s your angle.
I’m using the whitebox word very carefully, because there’s lots of rebranded assembly, and very few good management layers. If you want DIY, if that really will save you money, there’s always routes. Nexenta is certainly DIY. Somehow i doubt Intel will be unable to source chips or source licenses for its adapter range, which is to my the best of my knowlege LSI gear, with better support response. Nor, say, will Supermicro, whose onboard is also LSI usually, and it’s very good onboard. (muse for another day, quantity production critical storage hanging from a LSI chip on a Supermicro versus NetApp installed!)
Maybe i didn’t look into Nexenta enough. They’ve a lot of software which looks good, but on examination, really looks like what i can get from CA, running on a bunch more platforms.
Aside, here, please allow me self certificating a long history of saying very nasty things about backup s/w companies of all kinds. But lately Arcserve has been improving a lot, and functionally it’s not much different from what Nexenta would like to sell.
Yeah, i’m comparing Nexenta to a backup app shop, because the neat networking is largely Solaris kernel code. Nexenta also charge per storage unit, which is to me highly offputting, allowing i get no fully tested environment with the spining rust. I can put that money into support, and keep the change.
None of this strikes me as “Open” anything. Not that i have any great love, and i’m not suggesting a direct comparison, but 2008 Datacenter, the example chosen because MSFT is the default comparison for anything labelled “Open”, not because you raised this, really starts looking a bargain, in the context of standalone layered capacity licenses. Or even Oracle’s plan with Solaris.
First rule for me: cull layered recurring costs. First rule for anyone: pay as few people as possible. If you are not there playing with gifted budget, by which i mean you have to explain to a board your capital returns, fiscal end, as opposed to that being done for you, i’d argue take hits up front, and amortize yourself. It’s fun to haggle away recurring contracts inside a purchase .. hilarious fun if you like game theoretic footsie, and have an inking as to commission structures. (why else is this ex salesbody in my seat? 🙂
I digress, but what defines “open”? Toigo exposits a very thorough argument, but i can tell you that that well formed line is a real conversation killer. To me, open storage management would have to go down to the firmware level. But that’s the heartland of premium Q/A, and that is actually a premium I would like to retain the option of paying for. It’s my favourite bugaboo and rhetorical hobbyhorse, but T10DIFF is just a starting point, and i think the OEM parts which speak it fit on one hand.
Alessandro, I’m not bashing you here. Yours is a proper supply chain question to ask. I’m just uncertain the problem imagined touches so widely, or is in fact the problem. I don’t see one OEM value add integrator as gatekeeper, or at the very least i would want to study hard to understand whether the risk exists.
I’m more worried about true supply chain, in my buying cycle, as the USA and UK are having to mothball auto plants, and if that kind of problem is more widespread (check out NPR’s wonderful audio history of the NUMMI plant for cultural insights, link below) than the super JIT auto shops, i can tell you bigger outfits than mine are soaking up any inventory they can find. Capacity per platter growth will get stalled by the recent events, if not for technical reasons, but price elasticity. Yup, i’m actually pretty nervous, and simply can’t tell you how much i worry about macro things.
Please also forgive me a bias here. Nexenta was something i really wanted to use, but it fell on showstoppers which i don’t think detract from it practically for other environments. Solaris is also a old crush of mine. This comment was triggered by thinking “Err, Nexenta is DAS?” and wondering if they would like to be in that space, therefore the Engenio sale has frustrated them . . .
cheers all!
– john
p.s., nearly forgot, this is a good listen, and also a very poignant story of how poorly we all got it with manufacturing in such recent years (I really do think it’s a very recent screw up, both sides of the pond dumping on manufacturing) and might be sobering to note some stat comparisons between say California and Japan, population, land mass and output – wise. Too ashamed to put the UK in there, lest i get emotional.
http://www.thisamericanlife.org/radio-archives/episode/403/nummi
Just wanted to point out, this is nothing new for NetApp. They also “branched out” with their VTL product. Now, that may not be the best example as it’s now being shelved (for good reason), but they have diverged before. They were planning on replacing their VTL with DDUP, but we all know how that ended. While they primarily focus on ONTAP, they do know how to sell product outside of that scope.
Aren’t most tech acquisitions mostly about the tech than they are about market share, etc?
I am presuming this purchase includes the technology LSI acquired from it’s purchase of OnStor Inc?
Or did I miss something?
I am presuming this purchase includes the technology LSI acquired from it’s purchase of OnStor Inc?