CRN Exclusive: CEO Doron Kempel On Why SimpliVity Trumps Amazon, The Cisco Partnership And The Battle With Nutanix
Kempel On The Record
SimpliVity CEO Doron Kempel is taking direct aim at Amazon Web Services' public cloud economics as he charts the next stage of growth for the high-flying hyper-converged startup.
Kempel told CRN that the SimpliVity technology architecture offers data center customers a flat out better value than does public cloud market leader AWS. Midsize solution providers and enterprise IT executives have a compelling economic advantage using SimpliVity, he said, rather than Amazon because of the cumulative costs for VMware virtual machines over a three-year time frame.
Westborough, Mass.-based SimpliVity is making significant gains in the enterprise data center market with its Cisco OmniStack partnership, Kempel said. With its tight Cisco partnership and plans to offer its software on any server VARs or customers choose, he asserted, SimpliVity can undercut Amazon on price, and beat it on performance while locking up big margins and growth opportunities for partners.
Here is an excerpt from an exclusive CRN question-and-answer session with Kempel.
A lot of companies have tried to compete with Amazon, and a lot have failed. What do you have that's different?
The first thing you need to do is find technology that runs on x86. We've delivered that. Now, we can't allow you to buy the x86 at the same price Amazon does. But what you need to look at is we have developed a technology on top of x86 that runs better than the Amazon technology that was developed 10 years ago -- with deduplication, with compression, with protection, with all the intricacies around what enterprises want. This is what we've built this product to do. The assumption is that our technology on x86 is more efficient than what you would pay competing against Amazon. If you want to do that, you can't buy 10 different boxes; one of them does the server; one of them does the storage; one of them does the deduplication for backup; another one does the deduplication for WAN; another one does the deduplication for moving data to the cloud; another one does backup for VMs. ... We've developed all of that for that particular mission, and according to us, we've succeeded.
Where does the Amazon architecture fall down?
I think that [SimpliVity] changes the tenor of the dialogue, whether you develop your own private cloud, or maybe you're a VAR and you become a cloud provider with low-cost x86 running our software. SimpliVity developed a data virtualization platform, delivers next-generation convergence, cloud-like simplicity and economics, but also enterprise capability. All those capabilities are not capabilities that Amazon and Google try to deliver. What they run on their farm is very different architecture than what JP Morgan Chase runs on their system, where each VM is so important you're not allowed to fail. We developed a technology that allows even small players to achieve the economics I described.
How does SimpliVity position itself against Amazon? In the market as a whole?
I shouldn't be saying SimpliVity against [Amazon Web Services], but rather the alternatives to running your IT on AWS as opposed to running on-site on SimpliVity or running at a service provider on SimpliVity. If [a partner] wants to establish his own cloud, we want to convince him to use our technology. There's no third-party storage running on OmniCube. No Data Domain, no Riverbed. IT is complicated, and if you look at the top of the stack, you have orchestration and management, and below that, typically you have a bunch of servers, a switch and highly available storage. Typically when people talk about hyper-convergence, they talk about that area.
They talk about the hypervisor, say VMware, they talk about servers, the switch that goes inside the server, and storage. But the world spends a lot of money on SSD arrays, sometimes separately, and on Data Domain -- EMC bought Data Domain for $2 billion. This is a big business for EMC -- then comes Riverbed. Then there's cloud gateway for people who want to go to the cloud, storage caching and data protection. Deduplication and compression. What we suggest is that hyper-convergence is all of that.
Is all this academic because of the massive shift toward Amazon?
It's not academic, because if you're a [chief information officer], you're considering to run these types of applications on-site. The ones that require predictable performance, data privacy, governance and control. The cloud, what they tell you, is going to be subject to lower TCO [Total Cost of Ownership], elasticity, the ability to expand and contract, and agility. What we do is we allow you to grow very quickly. You grow like Legos with these very small building blocks. We have small, medium and large systems. So agility and scalability are part of what we offer. When you add more systems, there is no diminution of your performance.
What is the difference between SimpliVity and Nutanix?
Nutanix takes the top of the stack, the servers and storage, and builds an appliance in which the storage is designed to live inside the server. It's a virtual storage appliance, but instead of being just storage, they run business applications inside the server. However, it doesn't address everything. … You still have to buy Data Domain. You still have to buy Riverbed. What we're saying is everything below the hypervisor, all of that, becomes software that runs on x86, which is what we're offering today. We don't do anything above the hypervisor. We plug into whatever management and orchestration the end user chooses.
Nutanix gets a lot of buzz -- the founder's the guy who did the Google file system. Talk a little about SimpliVity's DNA.
There are a lot of excellent engineers here on the East Coast, as we all know, and in order to develop that technology, we needed a Noah's Ark version of what needs to be done. We needed people from VMware, we needed people that understand networking very well. We needed people who understand hardware very well because we had to develop our own card. We needed people who understand storage very well. And all of those capabilities need people who understand deduplication and compression very well. All of those are capabilities that sometimes you cannot find in any one company, and by the way, Amazon and Google have not developed the technology for the enterprise. What we're trying to do is combine the best of both worlds. We want to run on x86, which is very important, but we also want to be very sensitive to what it means to run Tier 1 applications, and we understand it requires a lot of resilience.
What is the difference between how SimpliVity does development versus Amazon, Google or Facebook?
When you develop a file system for Amazon, they probably have 500 people who do that. We have a lot of very talented people. Unfortunately some of the distributed computing work that we do we couldn't even find here on the East Coast. So we hired people [at] Cornell [University] because they have some people who understand the types of algorithms we use. The whole idea is this was complicated because we need a lot of multidisciplinary expertise that doesn't exist in any of the companies you just mentioned, and we need all of them in the same house combining what works for cloud providers and what works for the enterprise.
Talk about the SimpliVity architecture and how it was developed.
It all starts with having the right architecture. It's the technology. It needs to be part of your initial design, and then you need to dig very deep. You need to develop that object storage like we did and the file system. We said to ourselves we need to dedupe the data before it ever hits the disk. Amazon doesn't do that. Nobody does that.
When we suggested that idea in September 2009 -- remember, I was CEO and founder of a company called Diligent -- they said it's ludicrous. The engineers said you're crazy, you're going to slow down the data. But with the right technology, you don't slow down the data, so we developed the card. And this card crunches the data at inception, so all the data is deduped and compressed and optimized once and forever, globally.
How do other vendors stack up with regard to deduplication of data?
Every vendor you speak with is going to tell you "I dedupe too." You dedupe where? You dedupe at inception? Post-process? You do the backup, you do the WAN, you do the RAM, you do the flash, you do the ACD into the cloud. We do all of it. There's nothing left to dedupe, compress or optimize. I suggest that nobody [else] does that.
If you want to compete against this very capable company, you need to compete by offering a technology that was meant to compete against that reality. And you need to run it on x86.
What percentage of sales are going to be on x86 servers separate from the Simplivity OmniCube?
I would say in 2017, our fiscal 2017, which is coming up now, our expectation is that approximately half of our business is going to be our software on servers that are chosen by the end users or the VARs. We've invested very heavily in qualifying our software on those servers, and if customers want to consume the product in that way, we're ready to do that. We have a network of distributors who can basically deliver in that particular manner. It all depends on the demand of the end users and the VARs.
What kind of traction are you seeing with Cisco OmniStack?
It's been moving much faster than we thought. We went into 2015, I projected a single-digit number to account for our OmniStack sales -- OmniStack is our software on [x86] servers. We ended the year approximately 5X. So we have sold 5X what we anticipated. For example, our business with Cisco has increased three times Q2 over Q1, then three times Q3 over Q2 and then about two times Q4 over Q3, which was already a big number for us. The importance of the relationship with Cisco is measured in more than just dollars. Think about the type of partners we get to work with. Think about FusionStorm, a major player.
Cisco has a certain halo. The relationship goes well beyond dollars -- it has to do with getting the right attention, and not just the very large partners, but the very large deals.
When it comes to closing deals, how much does the relationship with Cisco help SimpliVity?
Deals that are significant to us that would have been more difficult to win without Cisco. In Q3, we won a very large deal with the government.
This quarter that's about to end, we won a deal in Europe with one of the largest companies in the world. Once you get into those particular accounts, you know there's tremendous room for growth, it's just breaking in that is the hard part.
In Q3 with Cisco, our largest customer ever is a service provider that decided to go all-SimpliVity, 200 systems. That is a service provider that is willing to run 100 percent of its workloads only on SimpliVity running on Cisco rack-mount servers. This is very meaningful to us. Could we have won the deal without them? Probably, but it makes everything much easier when we bring this large player, or they bring us, into a deal.
Talk about what has been one of the most notable Cisco deals.
Probably the most important customer we've had since we sold the first few systems in Q2 2013 is a global 100 company. That customer actually was in dialogue with one of our very important VARs, and it's a customer that prefers a different server vendor other than Cisco, and we were not qualified on that other vendor. We went into that deal with that VAR, and we won that particular deal. This is a deal that is going to include a complete technology refresh across all their data centers around the world. Hundreds of systems with a rollout plan for 2016.
This particular deal changes the narrative. That customer has chosen SimpliVity hyper-convergence to transform all their data centers. This is a big deal in terms of our relationship with Cisco. It means hyper-convergence is not limited.
Why isn't Cisco more vocal about the partnership with SimpliVity?
EMC acquired VMware. Did it make sense for everyone else to develop their own version of VMware? It was very clear that VMware was going to be very important. Why didn't Oracle develop their own version of VMware? Dell? Sun? IBM? Cisco? All of these companies should've developed their own version of VMware. Why didn't they? They couldn't do it, and by the way, it's late in the game.
If you're Cisco right now and you're looking at SimpliVity and you see all these large deals we're doing with them, you're probably asking yourself, "Is this going to be another VMware, and if they are, how are we in better control of what's going on?" Well, you can reach control by developing your own alternative, or by acquiring, but we're not really for sale.
Why isn't Cisco more communicative? First of all, we love the partnership, they don't want to talk about it, and that's OK. We just want the relationship, not the bragging rights. I think that will come because people will know about these deals we're doing together.
How is SimpliVity's relationship with Cisco affected by Cisco's investment in Springpath?
We understand that Cisco is investing in Springpath, and maybe Cisco will acquire Springpath, and that's fine. And when I say that's fine, I'm not dismissing anybody. I'm just saying there's significant traction with Cisco and I don't see a reason why that would diminish.
If you are a VAR, you want to win the deal. You want tight registration that protects you and you want high margin. And importantly, you want to be able to retire quota with your favorite server vendor. SimpliVity "Meet In The Channel" offers all of that. If Cisco OEMs a different product, it doesn't mean you get better value in terms of your ability to win, your ability to register, your ability to get high margin. Our win ratio in Q3 was 81 percent. That means we win 4 out of 5 qualified opportunities.
Is the Simplivity-Cisco relationship different now that Chuck Robbins has taken the helm at Cisco from John Chambers?
No. They've both been gentlemen. It's been easy to communicate with both of them. They've both been very responsive. They both shoot straight. It's been a pleasure.
Chuck has been super responsive. The relationship was envisioned in 2014, but it actually got on the road and required all the heavy lifting in the last two quarters. I don't want it to seem as if one of them did more than the other. Chuck is great. We've seen no change. We love working with him. He's very responsive. I can text message him. It's only good things about that company. They're great to the channel, they're great to work with, and I haven't seen any change since that transition.
Where do you see the Cisco SimpliVity OmniStack sales going this year?
Up and to the right. Business is going to grow, the number of partners is going to grow, hyper-convergence is getting more attention. SimpliVity is getting more attention. Everything is going to multiply in 2016.
What percentage of SimpliVity sales do you think will be Cisco OmniStack 2016 vs. 2015?
The company is growing at a very large speed irrespective of the Cisco portion. We've been growing 30 percent to 40 percent quarter over quarter. The company is growing very fast. Within that, if you just keep Cisco as a percentage of that, the numbers grow correspondingly. If we did $1 with Cisco and SimpliVity doubles its sales year over year or triples, that means Cisco is so going to benefit. Instead of $1, it'll be $3. The Cisco business is going to be very large, but not too large.
It hasn't slowed at all?
No. It's getting faster. Deals are larger. Thirty percent of the sales up to now in this quarter are from repeat customers. That tells you we're in very large enterprises and they're starting to buy more and more. That's an easier deal. It takes less effort. It means the sales force and the partners can work on other deals. Another very important variable is that partners are becoming self-sufficient. A year ago, we had to go in with each one to win the deals. Today, they do it by themselves. There's a multiplier effect. We're reaching a further inflection point.
Companies like Nutanix and Pure [Storage], when they got to their third and fourth years, they saw that inflection. Nutanix had a great inflection because VDI was such a great opportunity for them and they got into the federal market in their third year, which was terrific for them. We're getting into that particular zone. Remember that what we do is a little more complicated, because what we're saying is you can run everything, not just VDI, on x86. You choose the x86.
What is the message to VARs looking at the hyper-converged opportunity?
The message to VARs is we're not a software company, we're a partner company. Everything we just talked about was designed around the VARs delivering the value and being very successful. We keep the margin with them, they get to win the deals. They control the deal, basically. They get a lot of margin, they choose the server so they can keep their server partners happy. A lot of our VARs are becoming service providers. We have a lot of VAR partners that now have clouds. We've shipped over 4,000 systems. And that 4,000 systems counts where we are before we've finished the quarter. Some of these customers are among the largest in the world. They include one of the five largest media and entertainment companies in the world, one of the five largest oil and gas companies in the world, one of the five largest telcos in the world, and the U.S. government.
How important is the deal registration program for SimpliVity partners?
If you're buying directly from SimpliVity -- this is the Meet in the Channel program -- you register with us. Nobody else can go around you and get the discount we offer.
Since the software is a significant part of the deal, there's nobody else that can sell the SimpliVity solution to the end user other than the VAR who registered. Let's compare that with a case where you're buying on an OEM basis through a vendor. You may have registered with the vendor, but let's look at the Nutanix situation. Let's assume you've registered with Nutanix. They're trying to sell you their product. Dell can still go around that registration, and if you've registered with Lenovo, Dell can go around that registration. Or if you've registered with Dell, Lenovo can go around that registration. They have the exact same technology. If the customer is wise, the customer is going to call before the end of the quarter either Lenovo or Dell to see if they can get the same product at a better price.
Talk about what other server vendors you may look at bringing on board.
If you're in our shoes, you have to prioritize. Every time you introduce another server vendor, your other server vendors are less happy, so you just have to be clever in that regard. Eventually, any server, anywhere, running our software.
When it comes to EMC-Cisco vBlock, why is the SimpliVity path the better choice for VARs?
If everything I described to you is indeed true, then what does the world look like in 10 years? Is there any external storage, or does it all run on servers? You need to manage that transition. Everybody needs to manage that transition diplomatically. What we need to do is keep investing like there's no tomorrow in our development, to make this technology available on any server, under any hypervisor, under any management layer, and make sure our partners get the best technology with great registration and great support. This is what we're doing. It takes time.
Where do you think you're going to get the most VAR mind share?
The best VARs we're looking for happen to be partners of Cisco and happen to carry VMware and understand storage. They don't need to be trained. Cisco is very important for us, absolutely.