There’s simply no arguing with the numbers – EMC continues to do very well indeed in the broader storage market and continuing to gain share in almost all sub-segments.
While we’re not ones to rest on our laurels, I thought it would be good to dive down a bit deeper into some of the more interesting stories behind those big numbers.
For example, our mid-range product division (EMC’s USD – makers of the VNX and VNXe storage arrays) has certainly had an exceptional year. Many of the powerful forces at play in the industry happen much faster and stronger in this incredibly competitive segment of the broader storage marketplace.
I took some time to sit down with Eric Herzog, SVP of Product Management and Product Marketing, to get his views on 2012, and what things we might see in 2013.
We’re fortunate that there’s a lot going right for us now as we speak.
I think it all starts with customer adoption, which has been quite rewarding. What I find interesting is the incredible breadth and diversity – almost every industry, customer size, geography, etc.
Some storage vendors have to focus on only one or two niches; we get to work with just about everyone.
And – thankfully – they’ve told us globally that they’re very satisfied with what we’re doing for them. That's strong praise from a notoriously tough crowd. Of course, there’s room to do better, which we’re always working on.
Right behind that, I’m absolutely delighted with partner adoption: they’ve taken a very strong product, and added all sorts of incredible value around it to satisfy their customers – and their businesses are growing as a result. That’s great to see, and I hope it continues.
From a pure product perspective, we’re in a strong position as well. For example, my team focuses a lot on availability as seen by the customer, and we’re now moving beyond five 9s. Just to put that in perspective, that’s getting to be less than 5 and a half minutes of downtime per year on average. Very good, but there’s always room to do better.
The performance story – thanks to how we use flash – has been phenomenal.
The VNX is somewhat unique in that it can use the same flash drives as both a storage target or a persistent read/write cache. What the customer sees is simply amazing performance across Oracle, Exchange, SQLserver – just about any application that beats on data intensely.
The extended integration story has gotten simply overwhelming with VNX. It would take hours just to list off all the various integrations we’ve done.
As just one example, witness the success of VSPEX with our partner channel. We've clearly out-flexed the FlexPod. From a partner perspective, it’s beautiful – they get their choice of server, network and virtualization stack – all optimized around specific use cases. It’s almost like we’re custom-building products for our partners.
People are noticing what we’re doing as well, and we’re getting a lot of recognition as a result.
In 2012, we picked up 35 product awards, with 12 of them in just the last quarter.
For me, the most satisfying ones are the end user surveys – not just the storage users, but application users as well who point to VNX as the best-in-breed for what they need to get done.
The VNX products are doing quite well in what is perhaps the most competitive segment of the storage market: there are many products from many vendors. What do you see as the keys to success in this market going forward?
Well, let’s start with the basic premise: a unified storage array that does it all in one box, simply and efficiently with incredible power and performance. That's become table stakes in this market.
While the needs of the storage and infrastructure team are important, more and more storage is being directly used by application owners: Oracle DBAs, Exchange administrators and so on.
They want storage to appear as part of their extended workflow, and not have to go to the storage administrator to get anything done. Along these lines, we’re also seeing strong positive reaction to the application-specific tuning we’ve done right out-of-the-box. For example, if you’re an Oracle DBA, not only does the VNX speak your language, it knows exactly what you’re trying to get done, and configures itself appropriately. That sort of application integration and optimization is going to be even more important going forward.
If we circle back to the infrastructure team, they care first-and-foremost about availability – naturally -- but right behind that it’s optimization: cost-effective performance and cost-effective capacity – getting the most work from their assets as simply as possible.
I think you’ll also see customers increasingly value more integration with the other infrastructure components, whether that be the server, hypervisor, network, etc.
Whether that takes the form of a Vblock, VSPEX or something else, the story is still the same – customers value integration.
Finally, we build our arrays out of the same industry-standard components that everyone else uses, and that technology base moves incredibly fast. I think we’ve been able to do a strong job of turning around things like newer Intel chips, getting them into our products quickly, and doing it in such a way that’s as non-disruptive and investment-preserving as possible.
And we can only get better at that game going forward.
There’s been an enormous focus on integration with VNX: converged infrastructure, VMware, specific application environments, VSPEX, etc. – why is this so important, and what do you see coming down the road?
Storage is just one part of the overall environment, it’s something that everyone uses. And, generally speaking, more integration delivers more value -- for everyone, and not just the storage administrator.
For larger environments, the general trend is easy to spot, it’s modularization: standardized infrastructure bricks of compute, network and server. Whether that’s a VCE Vblock, a partner-led VSPEX, or something the customer is creating on their own, the VNX is turning out to be a great fit.
And a good part of the value we can bring is doing as much integration work as possible with the other components – hardware and software – that the customer has chosen.
Finally, it all gets down to management integration: making things easy to operate, day in and day out. That's not as easy as it sounds -- there’s a big variety of management models out there: storage-centric, infrastructure-centric, VM-centric, application-centric, service-delivery-centric.
VNX’s Unisphere management framework has done an exceptional job of supporting all of those different management models with a single, extensible set of capabilities that can be presented any way a customer wants to see them.
Some people might think that VNX is mostly a mid-market sort of product, but we’re also seeing a lot of adoption by larger enterprises – can you share what some of those use cases are, and what makes the VNX so attractive to them?
In the hands of customers, the VNX has developed very distinct personas.
First, as a shared block and file storage resource, it’s very popular – and those requirements show up in big environments as well as more modest ones.
Second, it’s an important building block in converged infrastructure. Every VSPEX is based on a VNX, and a surprising number of Vblocks as well.
But there’s also a rather unique role for the VNX, and that’s when you need to tune something to support a very specific application requirement – high-bandwidth, predictable OLTP, and so on.
In these spaces, you’re not really thinking about consolidating all sorts of workloads like you might do on a VMAX, you’re creating something for purpose, and you usually want to scale horizontally. Our customers can easily get into the guts of a VNX, and make it do exactly what they want for a very specific and demanding requirement.
Finally, no discussion around larger enterprises would be complete without a shout-out to EMC’s amazing Global Services organization: customer services, consulting, etc. – all the services and support they’ve come to expect from EMC. That’s a key part of what makes VNX so attractive to larger enterprises.
Just about everybody in the storage business realizes that there’s a key technology transition going on from disk to flash, and VNX has arguably had the richest feature set in integrating flash with disk at the array level. How is customer adoption going? Have we reached some sort of tipping point?
That’s certainly the case. Just about everyone has realized that just a little flash in an array can make a huge difference – with the right software of course to make it all efficient and easy to use.
And we’ve done a lot of things right here: FAST VP, our ability to use flash as both a storage target and a read/write cache, making pools and policies easy to set up and monitor, and so on.
But I think the second wave is starting to hit – there’s a growing awareness that flash in an array is more than just stellar performance, although that’s not to be discounted.
We’ve got all sorts of customer examples of serious capex savings: bringing far less hardware to the table to achieve a certain level of performance. There’s storage efficiencies as well, no need to short-stroke drives, or not use all your capacity because of “vendor performance recommendations” -- which I think is a weak sauce excuse for poor engineering.
Going farther, there’s opex efficiencies – far less need for system tuning, performance optimizations, laying out your data in specific ways, power and cooling savings, optimization of floor space, and so on. And I think that VNX customers know that if they’re hit with a new performance requirement, it’s pretty easy to achieve for not much money.
Put differently, what the VNX can do with flash is scale performance and capacity independently. Having been around the storage business since 1985, that’s a big deal for customers.
I’ve noticed a substantial number of solutions we’re delivering that I’d call “VNX plus” – VNX plus VPLEX, VNX plus RecoverPoint, VNX plus DataDomain and so on. What do you see?
The base capabilities of a VNX/VNXe are no slouch, but we’ve also have to be able to tackle some very demanding use cases out there. It’d be great if there was one magic box that could solve the world’s problems, but that’s not the case, no matter what some people will try and tell you.
As part of EMC, we’ve got access to the broadest and most successful storage portfolio in the industry, and we certainly want to bring that value to our customers and partners if they need it.
Take VPLEX for example – it does stretched active-active clusters with capabilities that far exceed any array-based solution, and does it in a storage-agnostic way. We can help our customers by doing all the integration work between VNX and VPLEX up front – easier to buy, easier to deploy, easier to manage, especially in virtualized environments.
Or, take EMC’s RecoverPoint, the current market leader in continuous replication. Once again, its capabilities far exceed what many array-based solutions can do, and it also does this in a storage-agnostic way.
We’ve made things far easier for our customers by doing all the work up front, and that too has been very successful. The same general story is true for DataDomain and backup.
Here at EMC, we’re very fortunate to have access to all these wonderful storage assets. We don't have to do a force-fit between what we've got and what the customer needs. We’ve also got the expertise to assess a customer’s situations, and provide very tailored recommendations that may or may not include some of these other products. It all depends on what the customer is trying to achieve.
You’re going to see even more of that sort of deep integration going forward.
A lot of us have been tracking developments in the OpenStack community, and I think VNX was the first EMC storage product to offer Cinder plug-ins. How do you see the growing importance of different cloud stacks?
The cloud stack market is in its early days, so it’s important to be as flexible as possible, and not limit customer choices. We’re storage for the cloud – any cloud – based either the technology choices customers are making now, or perhaps the ones they’ll be making later.
Once you get past deep VMware integration which we’ve been doing for a while, you move on to deep Microsoft integration, deep OpenStack integration, deep Citrix integration, and perhaps more over time. When the market is moving fast, flexibility is key.
With regards to OpenStack in particular, there’s already a sizable community of enterprise customers and service provider partners who are using VNX as part of their OpenStack-based cloud, and I’m sure we’ll see many more in the future.
Any final thoughts?
I think it all starts with a big “thank you” to our customers: thank you for considering us, thank you for selecting us, thank you for trusting us. We know it’s a competitive market, and you demand the best. We’ll continue to do everything in our power to continue to earn your trust and your business.
Next up would be a hearty thanks to our growing cadre of successful partners: you’ve bet your business on us, and we know just how important those decisions can be. Let us know what we need to be doing better, and we’ll be listening attentively.
Finally, a big thank-you to the EMC team: not just USD, but the people across EMC who come together each and every day to delight our customers and partners.
We couldn't have done it with out you. And we're certainly going to need your help again in 2013.
Comments