Exclusive Interview With George Teixeira, CEO, DataCore

  George S. Teixeira, 57, president, CEO and co-founder, DataCore Software Corp.
He co-founded the company and is CEO and president since its inception in 1998. Prior to that, he served in a number of executive management positions including WW VP of marketing and GM of the product business group at Encore Computer, where he was a team leader of OEM marketing and sales to Amdahl, IBM, and DEC. His work culminated in the $185 million sale of Encore’s storage control business to Sun Microsystems in 1997. He also held a number of senior management positions at the computer systems division of Gould Electronics. He is a native of Madeira, a Portuguese island off the coast of Africa. He earned an MBA at Florida Atlantic University, after joint B.S. degrees in Chemistry and Biology. Hobbies: “I still play some racquetball, I’ve done the standup paddleboard surfing. I like bicycle riding and the company.”

Now you embrace totally the three words software-defined storage (SDS) in your marketing approach following the buzz.
Our mission statement when DataCore was founded in 1998 was ‘software-driven storage and architecture’. And the truth is, if we had called it ‘software-defined’ then the industry would have called it ‘software-driven’, but we went ‘software-driven’ so the industry went ‘software defined’.

Why not anymore storage virtualization software or hypervisor?
For a long time we had to evangelize the idea of software. The industry wants to still buy and think in terms of hardware, the mindset has been hardware. We kept trying to find ways to associate with software. Storage virtualization was the correct term in the beginning but when we started storage virtualization as a term we meant what today they’re calling ‘software-defined’, meaning it could not be sold as part of the hardware. Instead all the hardware people hijacked the term. So all of a sudden IBM SVC, Hitachi, everybody used the term.
Then, a few years back, because of the success of VMware, I said “let’s start using storage hypervisor” so that we could at least talk to the people who were using VMware, Microsoft. But the term had a short life. The big change was early last year or before when EMC announces ViPR and they start talking software-defined storage. They said two big things: commodity and software. And then VMware comes along and announces Virtual SAN, again software.
The two biggest player, and EMC the number one storage company in the world, says the future is about commodity and software. All of a sudden, my business, the opportunity has gone up by 50%. Because people are looking up ViPR and go “what else is out there?” and they find us. But the funny part is that, these terms, they are crazy, we’ve been doing the same thing for 16 years.

But who invented the terms software-defined storage and who did the buzz?
The first that I know is VMware CTO Steve Herrod. He did a speech and immediately EMC started building on it. The two started using that. But ViPR to me was the turning point.

Which company was the first in SDS?
I would argue that we were in the market much earlier than most of the companies. IBM comes out after us. But it depends which generation, if you go to the earliest days I guess you would have to give IBM the first use.

Who are you main competitors in SDS?
The quick answer is Nexenta, if you’re doing open. But if you take a close look, it’s not a company that we run into that often. Usually Nexenta sells a bundle, with a system. We don’t really compete with them because we usually are high-performance.
I guess you could argue IBM SVC is probably the other one. But in reality most of our competition is EMC, Hitachi, NetApp. But I sell as much with them, versus competing. For instance many of our customers today use EMC but what they do is they take VNX plus DataCore, and it’s cheaper and better as a VPLEX. In some cases they’re competitors, in most cases we work with them.

DataCore was born in 1998 and you’ve been its president and CEO since the very first day. For how much longer?
I still find it exciting. What’s amazing now is that the vision we had is actually coming real because I don’t have to explain what it is anymore. It took three times longer than I thought, to be honest.

Looking at our figures, $117 million was invested in DataCore since its inception with the following financial rounds:
1998: $2 million
1998: $6 million (first round)
2000: $35 million (second round)
2001: $37 million (third round)
2004: $7 million (fourth round)
2005: MBO (60% employee owned)
2008: $35 million)
I think that’s correct. It didn’t change since 2008 because we’ve been making money.

Main shareholders ?
The employees first, and close friends. And the biggest outside are Insight Ventures, Updata Partners and Flagship Ventures. Intel and Hitachi all made investments, but in 2008 we bought some shares so their amount is very little at this point.

Which figures can you give me to have an idea of the weight of the company?
We’re close to 250 employees worldwide, in terms of sales this year we’re in the $50 to $60 million range.
The net income doesn’t matter as much this year because I’m making investments.

Profitable?
We’ve been cash profitable for 3 years.

Any plan for an IPO?
That is still the long-term plan yes, but I’m in no hurry because to create an IPO in the U.S you need to spend $7 to $10 million and I can use that money to build on the product or to invest right now.

But you can get much more in return…
The market is crazy, you saw Pure Storage, $3 billion. It’s insane. If this changes right now then the reason to do it is to sell the company because somebody offers a crazy number. Right now my feeling is, I run the company to be a good one, and in three years I could IPO, if somebody wants to pay me billions of dollars then I would sell.

Percentage of revenue between licenses and services?
Probably 25% for services. 75% for licenses.

Average price per license?
On average it’s probably around $15,000. The last three deals for Virtual SAN that we’ve done are all over 3/4 hundred thousand. Because one was Wipro, they do 6,000 virtual desktops using SANsymphony for their own use. Another one, I can’t say the name, is part of a large restaurant chain in the U.S.A, and they’re going do over a thousand restaurants using the Virtual SAN. Ikea is still big, 50 or more licenses. Sanofi in France too.

Being 100% indirect, who are your biggest resellers?
Distributors in France for example are Miel and Avnet. In the world it could be Ingram because they’re so big. But in Germany we have one called ADN Distribution GmbH. Half their business is Citrix and half is DataCore.

On which virtualization platform do you sell the most?
Our business is still 75% VMware. Microsoft Hyper-V keeps growing though, it’s getting quite a bit bigger.

On which hardware storage platforms?
On servers, n°1 is probably Dell, but Dell and HP are almost the same. HP probably is the n°1 in storage. But we do a lot of EMC, Dell. It’s such a mix. There is not such a big difference now.

Is Germany still your first market in revenue? Why?
Yes, it still is. We started with our investment there early and when it grew it became critical mass.
In the U.S I didn’t do a push until we came out with SANsymphony-V in 2011. That’s when we started selling more in the U.S. And a year and a half ago I hired a gentleman named Steve Houck, he’s our COO, he came from EMC and VMware. In the U.S we didn’t had the right channel model and he basically changed the way we do. We grew last year by 50% in the U.S and most of the deals were large enterprises.
We hired a lot of insight sales people. Because in the American market you get on the phone, you call people and then you create meetings and we give the deal to the partner after we’ve already sold the work. It’s a different model. In Germany our partners are very certified. France is in the middle, the French team does a lot of the selling, of pressing. In the U.S we have to do almost 90% of the sales. We’ve had to put up a lot more investments. Part of the difference is in Germany you can do 50% of the sale, in France 60/70%, but in the US we have to do 90%.

OEMs?
Indirectly yes. In Germany we’ve announced a Fujitsu storage appliance. They’re taking the Fujitsu fast box and they added DataCore. They’re selling it as a Fujitsu storage virtualization appliance. Right now it is only being sold in Germany but the plan is to sell across Europe and then worldwide.

Has DataCore ever sold hardware? Will you enter in this field to offer a complete solution?
Never. We prefer the model like with Fujitsu. We could offer a complete solution but the problem with the hardware is that the moment you sell it it’s already obsolete because the next version is out. We would rather work with partner who really understands the hardware, Fujitsu is an example but we also created a system builder program.
We have something like 50 companies that will take Dell, Supermicro, make a package and sell it as a combined package. We felt that was a better approach for us.

What is the roadmap for SANsymphony storage virtualization software?
The long-term vision is a thing we call ‘data anywhere’.
The idea is that we’ve been able to do a lot of things with SANsymphony V-10 especially where the data can move from being on the server side to the SAN, into the cloud. And we believe that there is so much different types of storage, and that company will have different types. And the funny part is that everybody talks convergence but if you look at what they do, they’re selling divergence.
Today everybody is creating isolated storage islands. But they only work within their stuff, our features work across all of them. Our view is, we are a unifier of all these isolated storage islands.

What about object storage?
It’s still early in the market. We’re doing things like OpenStack for the API. And we are monitoring it. Also, because we work with Microsoft anything that they do on object storage we can also work with. But it is not a primary focus because the market is still relatively small.

Dispersed storage networking?
The Virtual SAN allows us to do that. You could run us anywhere in the network and we can aggregate.
Most of the solution are doing file-level. It’s what I call scale-out. We’re not really scale-out. We do file share and because we work with the Microsoft scale-out file system we can build very large files but it’s not our primary business. Our primary business is block storage. The file share is typically a section that somebody wants to just do documents and things like that. If you’re trying to do database transaction, that’s what we do best.

Do you continue to refuse to enter into de-dupe?
We don’t really refuse. We do replication, we use compression in forms of de-dupe.
Secondly we have file share. It does de-dupe, it’s the same as the Windows file share. What we don’t do is block-level de-dupe. However we work with Simpana from CommVault, we work with Veeam, with all the backups, every one of those includes de-dupe.
We also de-dupe writes when we do like VDI. If we see the same writes to the same place, we coalesce the writes so that you only write once. We do variation of de-dupe. What we don’t do is block-level, production primary de-dupe.
We work with caching also. If I see write traffic I avoid writing to the flash. If I see the same writes I hold them and ship them through auto-tiering somewhere else. The bottom line is I get 10x better performance than anybody else. I know we do that kind of form and if you’re trying to use flash you have to really ask the question: How much de-dupe are you trying to do? Because we do a mix of all the different storage, so unless you live in a pure flash world I don’t know what is the advantage. If your goal is to save money you will save a lot more money using DataCore and get the same performance or better.

What is missing at DataCore to become a big storage company?
I think most of it is marketing. And that’s part of the reason I brought in a marketing person. In Europe we hired a German named Michel Portelli. So we added a whole new marketing team and if you look at our website it’s much more modern. 

Your annual salary?
I don’t really go into the personal, it’s good. I’m not the highest paid in my company, our sales people are because when we started the company we made a lot of investment and I always felt that the people who sell should be getting the most of it.