Interview With 3par’s CEO David Scott

Key Facts and Figures on 3par

Company:
3par, Inc. (name coming from three of the original founders Price, Ashok an Rogers)

HQ:
Fremont, CA

Offices:
Sales offices in USA, Canada, China, Germany, India, Japan, Korea, UK and Singapore; R&D in California and Northern Ireland

Founded in:
May 1999

Total investment:
$183 million in five financial rounds, main investors being VCs Mayfield, Worldview Technology and Menlo Ventures, but also Cisco, Sun, Oracle and Veritas

IPO in:
November 2007 with net proceeds being about $95 million

CEO:
David Scott since January 2001

Total compensation:  
$716,000 in 2009

Products:
Utility Storage solutions for large enterprises

No. of patents:
Around 30

Acquisitions:
None

R&D:
210 people, $35.3 million expenditures for the nine months ended December 31, 2009

Sources of components:
Hitachi GST, Seagate and Xyratex for HDDs, STEC for SSDs, Power-One for power supplies, Renesas for ASICs, AsteelFlash and Xyratex for controller nodes and chassis assembling

Market cap:
$579 million

≠ of employees:
627

≠ of customers:
"More than 400 one year ago", according to Scott; more than 100 in Europe

Biggest customers:
Crédit Suisse, MySpace, Savvis, US Department of Justice

OEMs:
None

Main competitors:
EMC, HDS, IBM, HP, NetApp, HP, Oracle/Sun, Dell

                
Key Financial Figures of  3par
                        (FY ended December 31)

  
Distribution
   of Revenues
     for 3Q10

About President and CEO David Scott

 
David Scott, 48, is an English citizen born in Jamaica. He joined 3par as president and CEO in January 2001. Scott has 26 years experience in the enterprise storage, software, and server market. He was previously GM of HP’s XP enterprise storage business in its Network Storage Solutions organization. Prior to this role, he held various marketing, business strategy, sales and systems engineering positions in HP’s HP9000 enterprise server, HP-UX software and enterprise storage businesses both in the US and the UK. He holds a Bachelor’s Degree with Honors in Computer Science and Mathematics from Bristol University in the UK. Hobbies: playing with his seven-year daughter, chess, skiing and reading.

Q&As

StorageNewsLetter: How many SSDs have you sold?
Scott: We have only just started to ship them, in fact we only announced a few weeks ago that we are starting to ship in this quarter…

But do you see this market as really fast-growing, because it seems to me that the number of sales is not huge…
Our view has always been that solid state disks have three potential uses. One is for low-latency applications, like trading locations, where they are better located in the server. The second is to overcome weaknesses in some legacy storage architectures, where you cannot get a very large number of I/Os into a single volume, which is why we believe many of the legacy storage vendors introduced solid state disks up to two years ago. We were able to address it within 3par Utility Storage, using wide striping as a mechanism for getting a lot of I/Os, a peak number of I/Os to be very high.

For cache?
No, we use it as a tier of storage. And the third application area for SSDs has always been cost-optimization in our view, and the only way that SSDs can be used for cost optimization is if you were able to deliver sub-volume level autonomic storage tiering, where the system itself could examine individual volumes of data, look with the regions of those volumes, look at the I/O density access patterns, and then move those highly accessed regions to the appropriate level of disk, which would include solid state. We introduced Adaptive Optimization, which is our implementation of autonomic storage tiering, at the same time as we introduced the SSD platform with the Mac 8 IOPS [SSD drive from Stec], because that unique combination we think opens up the opportunity for solid state disks really to be a clear value to customers in the market place.

What do you think of primary de-dupe, and will you have that in the future?
We’ve led the industry, we believe, in three critical areas: one is in multi-tenant clustering, the second is in autonomic management of storage systems, and the third is in the use of thin technologies for primary data storage. In the area of thin technologies, we include not just thin provisioning, but also the ability to perform fat-to-thin migration, the ability to keep environments thin over time using technology we introduced called Thin Persistence, to integrate with external file systems like Symantec’s file system using thin reclamation technologies, it’s the implementation of our hardware-based ASIC of parity calculations for RAID that allow very high-performance RAID-5 and RAID-6 that reduce the amount of overhead in the environment. It also is the introduction of our Autonomic Storage Tiering technology which allows you to achieve cost optimization relative to let’s say FC-only solutions when you’re using our Adaptive Optimization product with solid state disks and SATA drives. The area of the use of primary de-dupe is one that we’ve watched closely, but primary de-duplication is best focused on file stores, typically, that may have lots of user files, your PowerPoint files, there maybe lots of copies around, and that area of primary de-duplication is not an application focus area that 3par has as its primary target market. And so primary de-duplication has always been of less important than many of these other technology areas that we’ve provided leadership in. We do not believe that the area of primary de-duplication today aligns with the transactional database-oriented target markets that we focus on.

We could categorize you as one of a number of small companies that compete with giants. Like for example Compellent, which is a relatively small, albeit growing company.
Well Compellent has traditionally targeted small-to-medium sized business through a broad reseller channel, whereas 3par has typically targeted very large enterprises, big service providers and government organizations through a direct sales channel in the U.S. and internationally through partners. We both use virtualization technology, but the difference is our platform scales massively through clustering, rather than just being a dual-controller implementation like Compellent. So if you like, we’re at different ends of the platform spectrum, but with a common set of competitors. And I think that both companies have executed well in being able to take market share.

Isilon is in NAS, but also another competing with giants. What do you think of them?
Isilon as you say is in the NAS space, and again competing with a giant, they clearly brought some very strong technology capabilities there. They targeted a segment of the data growth that we believe is a more difficult environment to target and make money in, because many of the largest consumers of data storage for digital and video content, which is their market, they find it difficult to justify the high cost of storage solutions in that space. Because it’s not a tremendous technologically challenging space, you know you store data, and then you retrieve it over time, it’s not a transactional environment. And that has led a number of the largest companies in that space to actually build their own file systems, so if you look at a MySpace, or a Google or a FaceBook, they’ve ended up building their own file systems and deploying those solutions on cheap disks with integrated servers as an alternative to the Isilon approach. And with Isilon trying to make gross margin in the 50 points competing with generic serves that have 18 points of gross margin, it’s a very difficult business in which to make money.

BlueArc?
BlueArc appears to have had a strong position that it’s gained in the high-performance network-attached storage space. I think they’ve broadened their product line out. We tend not to see them competitively at all.

Panasas?
Panasas again is moving into a slightly more niched area of parallelized network attached storage solutions, with I think applicability into some of the high-end commercial data warehousing spaces.

I’ll finish with a company which is more… I suppose you compete with Pillar?
I don’t think we’ve ever substantially competed with Pillar.

No? They area in high-end NAS/SAN…
I don’t believe that they’re in the same market segments we’re in, we very rarely compete with Pillar in any solutions. I think they tend to target small-to-medium sized businesses, mainly through a mixed channel.

Do you see a future for them?
You know, I don’t have an Oracle to be able to tell you that...

Two more questions. Your company is growing fast…
We’re very happy because we grew 10% last year.

But how can you improve your slim profitability, net losses for all reported fiscal years and an accumulated deficit of $178 million? What will you change about your company to be more profitable?
I think that the figure you stated is slightly misleading, for the following reasons. Ever since we went public since November 2007, we’ve actually achieved profitability on a non-GAAP basis. And the only difference between a non-GAAP earnings per share results and GAAP earnings-per-share is employee stock options expense. So on a cash basis, we actually have been profitable ever since we’ve gone public in every quarter.

But how do you shareholders feel about that?
We’ve also generated free cash flow at a substantive level over the last three quarters. So for a measure of overall profitability, I think we’ve achieved the important step as a public company in showing that we can make money. The strategy that we’ve taken, however, is to invest as much as we can back into the business, because we have a nine billion dollar target addressable market, and today we have about a 2% market share. And for our investors, we think it’s in their best long-term interests if we are able to invest as much as we can, especially coming out of this recession, in enhancing our sales and marketing approach to be able to add capacity, to be able to regenerate very fast growth rates looking into the future. Over the past five or six years, the company’s grown at almost a 70% compound annual growth rate, we’ve slid to about 10% last year, when most of our competitors in the high-end storage market went down 20% and we were up 10, but now we want to reaccelerate that growth rate. And for a company like 3par in a very large target addressable market, most of the investors who are attracted to invest in 3par want to see us grow market share, revenue market share, as rapidly as possible. Because they know that if you are able to do that with a strong gross margin story, you will be able to generate earnings per share down the road. And that’s in fact that strategy that we’ve employed.

But if you look for example at Compellent, they are profitable. They are also growing very fast.
I can’t compare ourselves with a company that is going after a different market space, but one of the things that you end up finding is that if you start to early trying to ratchet up your earnings, you will start being measured on a company basis on a price-to-earnings measurement rather than on an enterprise value-to-sales measurements. And once you start to climb up that profitability curve, it constricts the investment rate that you can make into the business, and sub-optimizes the actual absolute growth that you can achieve. And we believe that’s important, in fact if you look historically, we’ve had a tremendous track record up until the last year of recession, of being able to improve our operating leverage steadily over time. And once we’ve regained the absolute growth rate targets that we’ve had, we will continue driving operating leverage improvement until we get to our target operating model.

What’s your roadmap?
I think the most important element of our most recent announcement is the Autonomic Storage Tiering solution that we call 3par Adaptive Optimization. We believe we will have established leadership in the high-end storage arrays segment in bringing this functionality to market. Compellent, to give them credit, introduced that to the lower end of the market, but we are the first to introduce it onto the high end enterprise storage arrays. And so we think that there is a lot of run-time and play available for expanding that solution set into the market over the next couple of years.

I have to bring up cloud, of course, everybody is talking about cloud. You’re not really considered a cloud storage company.
There is a difference between selling storage that is designed to be the basic infrastructure for cloud computing, and selling cloud storage. They’re actually two completely different concepts. We sell our Utility Storage platform that is designed to be the core platform for utility computing architectures that allow the deployment of cloud computing and public and private clouds. That is where we focus Utility Storage. There are entirely other software-oriented solution sets that are more targeted towards delivering cloud storage, and that is not an area that we have focused on.