What’s the VNXe scoop in a nutshell?
- VNXe was code-named “Neo”. IMO - an apt name (read in the details to see why)
- VNXe fits all the goodness of EMC’s Unified VNX family into a smaller, even simpler, even more affordable package.
- When we say affordable, we’re not kidding. The MSRP is $9499. For this class of storage (feature rich, unified block and NAS, and expandable to 96 and 120 SAS spindles), that’s a GREAT deal - 20% less expensive than it’s competitors – some of which are block only, some unified (we compare the smaller VNXe 3100 to the HP P2000i, IBM DS3512i, Dell MD3200i, NetApp FAS2020)
- VNXe uses a completely homegrown EMC innovation (C4LX and CSX) to virtualize, encapsulate whole kernels and other multiple high performance storage services into a tight, integrated package. We call the process of prepare the storage services for integration “refactoring”. This is hands down, one of the coolest things I’ve seen. Hint – there’s a lot more getting ready to get “refactored” to be integrated into the VNXe and VNX family over time.
- It uses shares features/functions with the larger VNX family (it is running the same software, “refactored”), including Unisphere, which is drawing a lot of very positive feedback, including a “one click” online support experience.
- We made Unisphere on the VNXe even MORE easy, and integrate it with a set of storage use cases very common for the SMB customer – Microsoft Exchange, VMware, Hyper-V.
- Software bundles (like the VNX) for things like snapshots, replication and others have been simplified into 3 software packs.
- You can even leverage EMC Replication Manager for application-integrated local and remote replicas.
What’s important for partners (on top of the list above which customers care about, partners have other things they find important too)
- get trained, able to sell and service in 3 days – for free.
- best of breed program and overall profitability.
- dramatically simplified “ease of doing business” with all the associated systems and process stuff that implies.
In my personal opinion, VNXe is the coolest of the cool things we’re launching today… sometimes it’s cool to be small :-)
Read on for more, including demos and details…
Ok – here’s more detail on the hardware…
There are two members of the VNXe family – the VNXe 3100 and the VNXe 3300.
The VNXe 3100:
The “skinned” look (the VNXe 3300 looks very similar, but is 3u):
The “naked” array:
A zoom in on the storage processor of the VNXe 3100:
The VNXe 3300:
The “naked array”
A zoom in on the storage processor of the VNXe 3300:
The spec table:
A few noticeable takeaways:
- Each uses very current Intel architectures. The VNXe 3300 has 8 Westmere cores, and 24GB of memory – that’s a solid chunk of horsepower for something so small.
- They share the modular port expandability as their bigger siblings in the VNX family – the VNXe 3300 will have the full complement of Ultraflex SLICs that EMC customers are used to (10GbE today, FCoE and FC over time).
- They are remarkably scalable for something in this band – up to 120 spindles on the VNXe 3300 – that’s up to 240TB. Wow.
- While out of the chute, they use 3.5” enclosures, you will eventually be able to use the same 2.5” 25 drive disk enclosures that are on the VNX.
- They are active-active.
- They vault to flash. No more vault drives.
- You can self-service everything.
Hey – how are we doing this in a simple 2 brain package? All the competitors say EMC doesn’t do this? :-)
But the big deal is ABSOLUTELY the software.
I’ve hinted in the past about the C4 codebase and CSX projects, and Steve Todd has a great blog post on it here.
I agree with Steve – these could very well be the biggest internal innovation in EMC in years. It has 40 patents.
Why? The answer is these two pictures:
For every superpower, there is a weakness. For every superman, there is a kryptonite.
Or, put in less “DC Comics” style – everyone’s greatest strength is simultaneously your greatest weakness. I truly believe this of me personally, of people in general, of technologies, of companies.
In EMC’s case, our strength (to me) is: breadth of portfolio. If it’s storage, and you need it, we’ve got it. If we don’t have it, but our customers need it – we will built it or buy it. We’ve built a culture that absorbs acquisitions well. Not perfect, but well.
And, these are in some situations (again, IMO) are our biggest weakess: breadth of portfolio. Where a solution could be done “good enough” in a simpler way, sometimes we’ve needed to add more unnecessarily to a customer solution in terms of hardware, in terms of different “things”.
So – how do you turn a weakness into a strength? You don’t fight it – you leverage it.
Now – there are two ways to tackle this problem.
The first way is to try to integrate into one giant “new” monolithic kernel and software stack. The downside of this is that it doesn’t work. What happens down this path is:
- you lose an incredible asset - time tested, proven and battle-hardened software stacks. In storage land, this is really important. It takes around 5 “non-compressible” years to make a solid block stack. It takes about 7 “non-compressible” years to make a solid filesystem/NAS stack. The time is for development, but also for beating it into incredible hardened state through real world deployments. Remember, this “maturing” is really important - when you have a “bad day” in storage land, it’s a REALLY bad day. As an example, while Isilon is new to EMC, they have been doing it for 10 years.
- Even if hypothetically you pull off the one time effort, you’re still hosed. Winning means losing. Why? At that point, you are completely gated by your ability to update and maintain the monolithic kernel and software stack. You have also limited your ability to innovate to ONLY organic innovation – because any development that’s going on outside what you do is not in your monolithic stack. If you want to leverage that external innovation, you will need a multi-year effort to merge it in. I’ll put this another way… In other words, you’re making the unbelievably arrogant assumption that you can innovate faster than the entire academic and startup community. Inevitably your rate of innovation slows, and you get passed. It’s not the big that eat the small, it’s the fast who eat the slow.
Does that first way sound familiar?
The second way is the route EMC took. You build a platform that can accept “containerized” (abstracted, encapsulated) functional chunks from all sorts of sources. You make it fast, very lightweight, very scalable. You make it able to leverage the massively multicore x86 trend. That’s what the C4 codebase does.
Now, the question is – are you saying this is running a “virtual storage appliance”. Well – in many ways, yes. There are problems with the “just run a full VM” (and several advantages – so ideally you want that too!).
- The first is that many storage stacks don’t perform well in VMs for various reasons. There’s a reason most VSAs are limited to 1000-2000 IOps (about 10 15K RPM spindles, or a fraction of a single SSD). That’s not good enough if your goal is to be able to “refactor” just about anything – including what would now be considered “very high end” stuff.
- Hardware dependencies. In storage stack land, sometimes hardware dependencies exist. Now… In some cases they have little to no hardware dependency – think of Atmos as an example – it’ll run on anything. In other cases, they expect to see fans, power supplies, disk state and more. Note, I’m not talking about exotic, non-commodity hardware dependencies. EMC is trying to avoid those like the plague (we really don’t think that you can out-innovate commodity x86 hardware).
- Sometimes you need/want the full OS stack, but in a lot of cases, you wouldn’t.
So – you would build a thing that COULD run a series of types of containers. You would want to be able to scale out (many instances). Some containers would be lightweight system calls. Some would use an “EMC standard lightwight” wrapper (CSX). Some would use a full-blown encapsulated OS.
In the end, this is a MASSIVE innovation which has let us accelerate internal EMC innovation, and more importantly share innovations across the company, and absorb innovations from all sources much faster.
This means that VNXe, is literally, in a single active-active system that can be as small as 2U, has the mature, re-factored FLARE, DART, and other EMC bits that are in the bigger VNX family members. Common look, feel, function.
Don’t misunderstand, they aren’t fully encapsulated OSes as VMs, but as CSX containers. It’s extremely efficient, very high performance. It can leverage all the power we can fit into a tiny form factor.
Does this sound like something scary and new? Well, here’s another amazing fact. C4 has been shipping in EMC products for longer than you would believe. The first PRODUCTION GA code shipped in Q4 2008. Ever since then, core EMC platforms have been using C4 to introduce more and more functionality. VNXe – aka “Neo”, is the first time, however, when it’s being applied to something approaching it’s true capabilities… Not just adding new features in a modular, containerized way, but as the heart and soul of the platform itself.
Another interesting factoid – in the Nordics, we have been testing the core ideas with customers and partners. The NX3e was not VNXe but a focused “beta”. We were doing new things, and it’s always a good idea to check your ideas – and you can’t “kinda” check :-) We wanted to test out not only core principles in the product software design, but most importantly get partner feedback on the business processes and ideas. As VNXe is a product where the partners, the channel, play a huge role – that thinking has been built in to everything.
We wanted to make it so that the ease of doing business with EMC was easier, and that the VNXe represented not only a one-time opportunity for the customer and the partner, but would be something that would be the basis for ongoing partner/customer direct interaction.
Now - in other markets, delivering unified storage capabilities using extra hardware bits is not a big deal. As you get bigger, the extra hardware in a VNX (more filesystem blades, more storage pool blades) is “A-OK”, because those customers ask for the following: simple, efficient, performant solutions at very high scale. The pressures to on cost-of-goods-sold, physical footprint and other things mean that VNXe is the right place for the C4 code base to see the full light of day first.
Now, all of this C4 codebase technology is mind-blowingly cool in lots of ways… but of course – we would need to have a unified management model to deal with this modular code model…
…and that’s the bigger story behind Unisphere.
Unisphere makes it much simpler, much easier for EMC to innovate down a modular management model and integrate in all sorts of technology, quickly and easily. Common look, feel, and function. We’ve done it with Recoverpoint, Atmos, and others… and we’re accelerating. More importantly than looking at our own navel, customer feedback has been overwhelmingly positive.
But anyone thinking we’re stopping here is smoking something. And to anyone who says “EMC doesn’t innovate”, I say – “how do you like dem apples?” :-)
In all seriousness, at EMC, we innovate organically AND we innovate in-organcially. We don’t view those as two things, but a single thing – leveraging innovation. The C4 codebase and Unisphere help us do that. VNXe is a great example of what that means for customers.
A great quote from a respected (sincerely) competitor is an interesting way to close out:
“That said, over the years I think it's safe to say that the success of NetApp has been tied directly to ONTAP. Let's face it, if it hasn't been ONTAP-based our sales force has struggled to sell it. ONTAP innovation continued in parallel with convergence and that puts a lot of pressure on management interfaces. NetApp has made some big bets on improving systems management tools and must execute. EMC has strong management tools and an assortment of storage technologies. Convergence for EMC will take them down a different path (it has to if they want to compete in a timely fashion) but they must execute on integrating/converging what's behind the curtain. NetApp needs to work on the curtain.
Ultimately, I don't think the customer cares all that much - it can be hamster wheels turning SAS drives - as long as when they turn the storage apps up that it's all easy to manage and it doesn't slow them down. Execution - their business execution - is also the issue. The question will really be which is easier to do: pull together a management framework or pull together various storage apps behind a management framework.”
I agree with Mike – customers don’t care too much how we do it.
So - I’ve just shown how we’re both innovating around an integrated management model, and converging our core capabilities and stacks WITHOUT being tied to a single monolithic kernel approach. It let’s us converge today where needed, and converge future organic and in-organic innovations.
Oh – this ain’t just talk. We just launched it.
It also highlights why I keep saying – competitors shouldn’t focus on the other guy too much – we should each focus on the customer, and focus on ourselves, our own execution. It’s so easy to be wrong about the other guy.
But anyone – any competitor - who says EMC is just putting a “management wrapper” around stuff has NO IDEA WHAT THEY ARE TALKING ABOUT. (BTW, I don’t think Mike was saying that – of course, I don’t want to put words in his mouth – it is something I hear sometimes from people who don’t know better). It’s inevitable when you’re the leader – which EMC is – that people will be downers, I guess.
I’m unbelievably pumped about VNXe – for technical reasons, for business reasons, for innovation reasons.
Congrats to Doug Wood and the full team behind it, my hat’s off to you!
Here’s a set of demonstrations so you can see the incredible coolness:
- and really, really inexpensive (less than $10K for real-world configs)
A whole new world for us, and our customers.
Introduction (4:31) – A brief overview of the VNXe product family and a preview of the use cases that will be covered in the rest of the videos in the series.
Installing the VNXe (9:41) – This video shows the steps required to take a VNXe platform from power up to storage provisioning in 10 minutes and show cases the VNXe’s ease of installation features.
VMware Integration and Provisioning (11:50) – This demonstration explores the integration between the VNXe and VMware and includes provisioning of VMware NFS and VMFS datastores as well as the migration of live virtual machines.
Windows File Sharing (9:43) – Consolidating your Windows file shares onto a highly available platform like the VNXe is easier than you think. This video demonstrates windows shared folder provisioning and Active Directory integration with the VNXe.
Provisioning Microsoft Exchange (10:10) – The VNXe makes provisioning storage for your quick and easy. This video shows all the steps required to provision storage for Exchange and have it running in 10 minutes.
Data Protection Part 1: Snapshots and Local Replicas (9:57) – The VNX includes a number of features to protect your critical data. In the first of a two part video we’ll explore data protection using snapshots and local replicas.
Data Protection Part 2: Application Consistency and Remote Replicas (12:27) – To streamline data recovery and maintain business continuity, Application Consistent snapshots and replicas are critical as well as the ability to replicate your data at a remote site. This video shows you how with the VNXe.
Unisphere Guided Tour (11:46) – The Unisphere management interface makes the management of EMC’s unified storage platform simple and intuitive for application and server administrators even if they have never used shared storage in the past. In this video we’ll take a guided tour of the interface and see how the VNXe is actively integrated with the EMC Online Ecosystem.
So – what do you think? Cool, or are we smoking something? :-) Courteous comments always welcome!