what does that mean
Is the EQL tech in this system basically invisible to the end user? Does it do the same sort of "scale out" over iSCSI that EQL does? (I assume not).
With the worldwide availability of the SC4020 branded as Dell Storage, the company signals the coming together of its separate EqualLogic and Compellent networked storage array lines into a single brand and product technology. Simultaneously, Dell is OEM'ing to technologies for server-attached storage. We understand that …
Not sure how I feel about this - they're both just different products. One is an 'all-in-one', scale-out focussed SAN and the other is a more traditional, controller + storage SAN.
Ideally, the result will be a single platform with three types of box:
> Controller (SC8000)
> Storage (SC2xx)
> All-in-one (SC4xxx)
All three can integrate, tier, scale-up and scale-out. You will have the option of just buying multiple SC4000 units that, as they contain both the controller(s) and the storage, can scale-out just as the EqualLogics do.
Or, you can built a full, traditional SAN with separate controller and storage units.
Ideally . . .
The key will be the ability to mix-and-match and manage all together. For example, starting with a single SC4xxx unit, the way smaller setups do today with the EQLs. As you grow, you find that you require more storage so you add an SC2xx unit - scaling up.
As the SC4000 unit gets a bit older, you decide to add an SC8xxx dedicated controller, and a second SC2xx storage unit and then decom the original SC4xxx all-in-one box.
That will be great because the buy-in for a proper Compellent system can be a bit high compared to the one-box EQL. When it comes time to expand or update, those smaller installs will stay on EQL. With a converged architecture, they could proceed as above and move, gradually, to a full system. Or not.
For some people 2 x EQL is a great option so it remains to be seen if 2 x SC4xxx will be as good.
Of course, the EQLs run on ARM (Broadcom I think) while the Compellents run on commodity general-purpose x86. Are they really re-writing all that code?
Negative, this is a Compellent product rebranded as "Dell Storage" and parachuted into the EQL price range. It doesn't scale out, only scale up with another shelf (i think?).
Great news if you love Compellent. Otherwise it's time to evaluate the market as this new product provides an inflection point of change to any current EQL customer (just like Netapp did with the forced cDOT adoption to their customer base).
Well, that's what I figured, to be honest. There's simply no way you can 'converge' an ARM and an x86 platform when the only thing the two units really have in common is branding.
At least not without far more work than a recently-privatised Dell are going to be prepared to invest, which is a pity as EqualLogic really does offer something quite compelling in the market.
SANs are expensive so getting the most out of them is important. The expertise to spec, configure and maintain a SAN to achieve optimal performance is, however, not always available to a smaller provider. What that segment needs is a reliable, major-label (as their customers do ask) solution that is easy to configure and scale.
I'm not a Dell fanboy as my usual rut is HP but the product EqualLogic made and that Dell bought is a good one with a good blend of features for a good price. You can absolutely get better performance out of a Compellent but not without spending a good amount more money - either in purchasing the platform or hiring extra staff.
(BTW - Trevor: if you're reading this, I think I now understand your point about SimpliVIty.)
As a former Dell SE, who sold and supported EQL for many years, I can only say it was a great story that turned out to have a very sad ending for many customers. While the "scale out" concept works great on a whiteboard, having to pay for another entire array (controllers, software, and disk) every time you just want to add some storage, caused a lot of problems for customers - many of them initally attracted by the low intro price ('cause like anything Dell, the only sales motion the AEs know is "I can get that cheaper") ended up stranded because they couldn't affford to expand the EQL footprint.
Or they bought the sales pitch and mixed models (PS4000 and PS600) and ended up with performance problems because the 4000 couldn't keep up with the faster 6000.
And then there were the almost constant bad firmware releases... We never knew what they were doing with development (Dell was never very good about being honest with its employees about quality issues, let alone customers), but it certainly seemed that it was just a bunch of monkeys banging out code, with no test or QA effort at all. Don't know if they have managed to stabilize that, but I doubt it, because you would expect all development efforts to now be focused on CML.
It is absurd how complex many vendors products are, when you should be able to build much better intelligence into the software, but for many people complexity is career safety (more than a week of training just to learn how to operate an "Enterprise" SAN like NetApp (or EMC/IBM/whatever SAN ? Seems beyond absurd to me).
Nope, it's a dual core ARM chipset, exactly as it was all those years ago when launched. One core runs the network stack, the other runs the storage stack. This is why converging CML & EQL is near on impossible as the controllers are grossly underpowered vs the CML's multi-core x86 platform, and why this new SC model signals the death toll for EQL (sadly). Whilst Dell Marketing may not admit this, it's what the internal feeling inside the company is last I heard.
Even if Dell were contemplating refactoring the Equallogic code on x86 the underlying architectures are just too different for them to be merged in any meaningful way.
HP have a similar split in StoreVirtual (Lefhand) and 3PAR StoreServ. But StoreVirtual with Network Raid can do some clever stuff which would be just too expensive to attempt on 3PAR in the lower end of the market, that's why it still has a place in the portfolio.
I'm not sure Equallogic can really add anything that a low end Compellent can't already provide, and the fact Compellent supports standard Dell commodity hardware will make life easier and more margin rich for Dell in the long run.
This post has been deleted by its author
Hello. I am a director with Dell Storage product management. We expect that some commenters posting here work for competitors (such as Nick Dyer), so please take little heed to the distractors.
To be clear, the continuation of our strategy not only supports both current Compellent and EqualLogic customers, it’s going to make these offerings better in many ways and allow Dell to bring new features to market faster with a single Dell Storage architecture. With a common infrastructure, the future benefits customers using both Compellent and EqualLogic today. Specifically, we’ll be able to eliminate islands of data and offer a common way to configure, manage and provide application integration. And, our plans are to offer both scale out and scale up, offering customers greater flexibility that fits their preferred approach. The processor comments are irrelevant, as the new platform will work in concert with legacy products.
While this will be a multi-year evolution, we will first bring together today’s storage infrastructures with native replication capabilities that will go both ways. As shared with hundreds of our customers at Dell User Forum last week, this will be followed by common management and other on-ramps to the unified architecture. In the meantime, we still have plans to deliver new hardware and software updates for both SC-series (Compellent) and PS-series (EqualLogic) that will bring new features and value to customers along the way. To be clear, this is very much an evolution. Current Dell Storage products will be compatible with future Dell Storage products. Our clear plan is to ensure that no one will get “stranded” on legacy equipment.
As far as EqualLogic, in just the past few months, we began shipping our latest EqualLogic arrays, the PS6210 Series, based on significant investments. This includes a next generation controller and firmware that’s helping customers get notably more performance with flash compared to our previous arrays. Our software engineers are already working on the next two platform updates for EqualLogic with new features such as VVOLS support, compression and other enhancements already planned for release over the next couple of years.
Our PS Series arrays are the best solution for customers who need to expand their existing EqualLogic storage environment. The core value proposition of EqualLogic – ease of management, all inclusive licensing and scale-out architecture – remain unchanged, and these arrays will continue to be excellent solutions for customers seeking these features.
Thanks for calling me out directly. However I think you'll find that I was posting factual information on your new product rather than competitive FUD. In fact, I'm a multi-year Ex Dell Engineer (2 spend directly with the EqualLogic range). No matter which way marketing teams try to spin it, introducing a new platform does create an inflection point for customers. The great thing for us all in the market is innovation is rife and gives customers more choice than the usual EMC & Netapp, which is only good for the industry.
-- Disclosure NetApp Employee, Opinions are my own --
Firstly, I'd have to agree with Nick, that major product transitions typically do create inflection points.
I do take exception with @Nick using the transition between 7Mode and Clustered data ONTA as an example. Saying this is being "forced" on customers falls under the non-Factual form of FUD. The most recent versions of ONTAP 8.2.1 which is less than three months old now, includes numerous new features for 7Mode as well as support for all the most recent hardware releases, and has a long support timeframe. Suggesting that this is a forced transition makes it sound like NetApp has announced end of life, or end of support for 7Mode, and that is the worst kind of FUD, unfactual, misleading, and bordering on dishonest.
To be sure there are some differences between 7Mode and Clustered Data ONTAP from a management perspective, but they're more the kinds of differences you get when moving from one Linux distro to another, almost all of your existing knowledge about 7Mode is easily transferred to cDOT, and most people will choose the make the transition at the same time they're upgrading their storage controller hardware, so this is about as big a transition as any typical hardware refresh. ie. not much of one at all.
Other inflection points are similarly smallish, much like every EMC hardware upgrade, (e.g. VNX->VNX2), e.g. replicate the data, learn a few new tweaks on the management interfaces, and learn how to get value out of some cool new features, though unlike the EMC upgrades I'm aware of, at least with NetApp you get to keep using the disk shelves and drives you had on your old system.
There are however some much larger inflection points, usually involving big changes in the way the new systems are provisioned managed, e.g. HP EVA -> 3PAR. Even when the replication piece is smooth (which can be harder than it looks when you want to do things like preserve snapshots , WWNNs, IQNs, and policy settings), having two different storage architectures with different management domains and methods can be a challenging problem to solve, and usually increases the TCO. If a customer is willing to do this for one vendor, then they're probably willing to look at doing this for two (or possibly more), and that can create a significant opportunity for other vendors.
I don't know if Dell are unifying the management and replication of their entire storage product line including FC, iSCSI, File, Backup, and Object storage, if so, good on them, but unless Dell are truly committed to doing it, then my opinion is, that they may have bitten off more than they could chew. I would also imagine that it would be hard to be committed to your own storage strategy when you are also committed to making VSAN work, or Nutanix .. It makes me think about what a certain Mr Lincoln said about a divided houses, or another great teacher about having more than one master ... its not exactly a recipe for long term success.
"Suggesting that this is a forced transition makes it sound like NetApp has announced end of life, or end of support for 7Mode, and that is the worst kind of FUD, unfactual, misleading, and bordering on dishonest"
Like when NetApp draw comparisons against Gartner/IDS leading scale out NAS platforms comparing a single node to an entire RAID system. Or that time everyone was told how good QoS addition to OnTap will be when it's really just to try and balance the ridiculous amount of 'hotspots' generated by a aging and decidedly out-dated architecture. Do you remember when NetApp told everyone that they get 1.5 million ops per second? Each filer was limited to 10TBs of capacity when they can handle 5PB. Real-world? I think not.
Let's not talk FUD when NetApp is the reigning FUD world champion. Pot, kettle, black, something something...
Biting the hand that feeds IT © 1998–2020