Where has Dell's storage business model gone awry? Dell storage has exhibited declining sales for about two years despite Dell buying some of the hottest storage startups around, like EqualLogic and Compellent. Why is this? When Dell buys companies it goes to great lengths to keep their employees and ensure their culture stays …
Perhaps the problem ....
... is more fundamental.
Storage often isn't as "simple" as an organisation the size of Dell would like it to be (the "Simplification bandwagon"). When you arrive at a client, their current mess, and you wouldn't be there if there wasn't one, needs a careful guiding hand to make enough sense of it, to understand which bits can be effectively and profitably simplified.
By dumbing the exercise down to the point where cheap storage hardware margin is expected to cover the cost of the clever analysis, the business model starts to fall apart, and the quality delivered to the end client falls. This affects market perception and the value of any "consumed" vendor that used to be a specialist in a unique area.
Infrastructure is becoming a commodity, but the complexity of implementing the commodity bag of spanners for a specific customer's business process, isn't changing. Too many customers forget to build in appropriate billing engines to the infrastructure deployment, then realise after the project that they have no idea how much it is actually costing, and therefore what they're saving on OPEX - but it is VERY dynamic and shiny. Duh.
Slightly off-topic, but...
I've been wondering for a while whether something like Exchange (or any IMAP server, as well) actually stores a file sent to multiple recipients (especially with internal-to-internal emails) more than once. If it does, that's ridiculous. Surely pointing at a blob with a reference count would be much more efficient?
Comments? I'm just curious, really.
Re: Slightly off-topic, but...
Single instance storage is no longer a part of Exchange as of the 2010 release.
I generally configure a third-party archiving solution to grab anything over 13 months old, most of them offer deduplication functionality.
It's a bit of a pain in the butt and we still lose capacity/efficiency, but at least it forced an archiving solution on us (execs wanted us to "just use Exchange and tell people not to delete stuff").
Minor HP Nit
Minor nit re HP "one with a paralysed mid-range due to perceived EVA end-of-life and no 3PAR replacement products". The recent StoreServ (awful name) announcements have remedied this well, even if they'll probably "rightsource" development and support to Kreplachistan to save a few bucks.
Dell is in a trap that many acquirers suffer from. They have a lot of overlap between the product lines. This makes telling a strong story almost impossible. How can you convince a customer of your value proposition if you can't figure it out yourself? Dell needs to rationalize its story, and if that means killing off some programs, so be it.
But there is an industry wide problem here too. It's clear that storage is far more complex than even two years ago. Punters are asking the question, "Why should I buy expensive compression and deduplication features to save disk drive space when drives cost $100 per terabyte?" The pricing of cloud storage makes that question even more pressing.
There is a risk that, in the rush to out-feature each other, the Big Iron of storage will miss the market, which could form around open source LINUX offerings. It may just be a matter of time.
The title is a bit misleading. There is some interesting things going on in the Dell world, just not ground-breaking stuff.
As an EqualLogic customer (branch offices) looking at Compellent for our head office SAN, we're very much interested in the cross-platform replication and single pane of glass management option (replacing SAN HQ and Enterprise Manager) in the hopper. Storage Center 6.3 is largely re-written and delivers much higher performance, especially with the SC8000 controllers.
AppAssure was a solid buy. It's cheap and I honestly prefer the guest-based VM backup instead of host-based that we've been stuck with for a while. I've had a number of host backup jobs fail due to VSS issues on a single guest, I prefer being able to pause individual jobs and work on specific problem servers without kicking off another round of backups or what have you.
The DR4000 appliance is neat, especially for medium businesses and small enterprises. Based on the Ocarina technology, it gets pretty good storage efficiency and makes a good backup target. A new version is coming with DAS for capacity expansion.
Dell also has clusterable FS-series NAS heads for their block storage, and standalone NAS in the PowerVault NX line.
I think the real problem is one of optics, as this clearly shows. Dell has a lot of storage offerings and things coming down the pipe. If the author had sat down with a Dell rep beforehand this would have been an extremely short article.
Writing off either Dell or HP at this point is foolish, they still have cash and solid products on offer. We're probably not going to see huge groundbreaking things anything soon from either but both are working to provide comprehensive storage solutions. If they've got holes in their storage portfolios, you can bet they're trying to fill them with something that's at least as good as what's out there right now.
"[..] one with a paralysed mid-range [..] EVA end-of-life and no 3PAR replacement [..]"
What is this supposed to mean ? Perhaps you missed the largest 3PAR announcement for mid range that they have made in at least the last six years(that was made less than a week ago) ? I wrote a 6,000-word blog post on the topic myself covering every angle. The big press splash only covered a fraction of what made the big headlines.
Myself as a customer for those past six years have never been as excited as I am now about the future of 3PAR storage.
IBRIX and Lefthand are write offs for the most part, I wish 3PAR had ponied up for Exanet when they had the chance (they were in the running early on but decided against it).
Once EVA is buried, Lefthand is next - I expect HP to keep the VSA stuff around, but there's no point in keeping the hardware around when the 7000-series 3PAR is priced where it's priced at today. Even though lefthand hardware is nothing other than Proliant servers, I suspect they can slash the development costs by reducing the scope to VSA-only, and devote the rest of the R&D to 3PAR.
Compellent is supposed to have some good stuff coming soon.. they wanted to come visit to tell me just haven't had that meeting yet (though now with the 7000, I can't imagine anything they could come up with that would match it).
Re: "[..] one with a paralysed mid-range [..] EVA end-of-life and no 3PAR replacement [..]"
Compellent have had something good just around the corner for as long as I can remember, still hasn't arrived.
Seems odd that the 2 things you point out as HP's biggest problems are the 2 storage areas they actually have the best technology on the market - StoreServ and StoreOnce are the best mid-range array and the best de-duplicating backup technology around. HP are missing credible NAS offerings - both general purpose and scale out.
History is a great teacher
Dell's storage division seemed to go into cardiac arrest the day they announced their split from EMC. The patient doesn't seem to be doing much better today. How would I fix it? Absolve all those aquisitions and go back to OEM'ing best-of-breed products. Dell's roots of greatness are in vendor negotiation and suppy-chain distribution. Shouldn't they stick to what they do best?
Larry @ NetApp
Re: History is a great teacher
Larry, that doesn't explain how Dell has built a small Equallogic sales runrate into a $1billion a a year business.
Re: History is a great teacher
Did they need to buy Equallogic to get that run rate? Me thinks not.
Re: History is a great teacher
Don't know - unlikely, considering the lack of global reach that Equallogic had prior to the acquisition in comparison to Dell's customer base, but the margins on owned product are 2-3 times the margins on selling someone else's OEM gear,
Re: History is a great teacher
2-3X margins for making vs OEM'ing? Really? If that were the case, Dell would be making their own tape drives, routers, keyboards, desks and staplers. Simply not true. Dell would have given Equallogic global reach with or without buying them. Darren Thomas tried his best but unfortunately leaves a legacy of buying expensive storage companies that contribute continual declining revenue for Dell.
Okay, I'll bite, Chris
- Dell is nowhere in scale-out NAS
...FS line of products on market for less than a year...but growing fast. Integrated in all three primary lines (PowerVault, Equallogic and Compellent). 1+Petabyte filesystem large enough for you?
- Dell does not have a high-end array
...correct, but this market is small and not growing, The sweet spot is entry level and mid-range, which is growing. Dell's focused on addressing the largest part of the business and can scale where required into most areas where design considerations allow.
- Dell does not have an all-flash array
...Dell's on its second generation all-flash Equallogic array. There's all flash, hybrid and high density hybrid arrays, and that's just Equallogic. Dell's also got flash options in the PowerVault and Compellent line, as well as doing ACTUAL innovation such as flash in the servers connected on the PCI bus (how it should be properly engineered for applications). What are you talking about?
- Compellent is not setting the world on fire anymore
...last time I checked with a recent ex-EMC VNX customer, they had just purchased their 4th Compellent array in 6 months and couldn't be happier. Have you spoken to a Compellent customer recently?
- Dell does not have a scale-out SAN
...scale-out to what? or scale-out in what way?
Equallogic will allow you to scale out to 2+ petabytes should you choose to grow that much.
- Ocarina data reduction isn't cutting it
This is still largely a Data Domain market with everyone else getting a smaller piece of the pie, but the Dell line is strong, delivers what it's designed to and those customers that I've spoken with that have deployed it are content.
- Dell object storage is making little progress
...Object storage in general is making little progress. Is EMC's Centera space growing or declining? Dell's product's good, but niche for now due to multi-year application certification. Additionally, a lot of vendors want to move away from developing OO interfaces and just address to a file system, which is why Dell has sold its archive platform with a filesystem front-end.
- Dell does not have a good basic filer
...A good basic file would be - what? Windows Storage Server - check.
Something that scales on some good basic enterprise storage, like the FS3500 - check
Something that scales on a fully-functional SAN that allows the user to manage both the file and block through the same single pane of glass like the Equallogic-FS7600? Check
- Dell does not have a solid server flash cache product
...there was this guy called Chris Mellor who wrote an article in June of this year.
You should speak with him http://www.theregister.co.uk/2012/06/18/dell_hermes/
Or you could do your research around Dell's Cachecade designs, where they've innovated a server flash cache product for certain form factors (R820, R920) which mounts at the front straight onto the PCI-bus and is hot-swappable.
Yes, I work for Dell, so AC.
Re: Okay, I'll bite, Chris
Dear AC from Dell,
Ex VNX customer? that's pretty short term then as VNX hasn't even been on market for 2 years yet.
Re: Okay, I'll bite, as well
Well, er yes the VNX product name may be new and you may refresh the hardware every couple of years, but lets be honest, the underlying components Clariion & Celerra have been on the market 20 years now ;-)
Time stands still for no-one
Read a good article somewhere recently that traditional major players in storage were suffering given newer and more versatile entrants winning market share.