Exclusive Interview With PLX (Oxford) CEO Ralph Schmitt

Key Facts and Figures of PLX

Company:
PLX Technology, Inc. (name coming from Programmable Logic eXecution)
HQ:
Sunnyvale, CA
Offices:
Abingdon, UK; Shenzhen, PRC; Shanghai, PRC; Taipei, Taiwan; Yokohama, Japan; Seoul, Korea
Founded in:
May 1986
Became public in:
1997
CEO:
Ralph Schmitt since November 2008
Total compensation in 2009:
$350,000
Products:
ICs for system connectivity: PCIe switches and bridges, DAS and NAS SOC and controllers, I/O and USB controllers
Main fab subcontractors:
TSMC
≠ of patents:
26
Acquisitions:
Oxford Semiconductor in January 2009 for 16.4 million, NetChip Technology in 2004, HiNT in 2003
Sales:
$82.8 million in 2009 (+2% compared to 2008) – Oxford contributing for $25.7 million for the year ended December 31, 2009 – , $118 million predicted by analysts in 2010
2009 net loss:  
$18.8 million, ($56.5 million in 2008)
Sales per country in 2009:
China 35%, USA 16%, Singapore 14%, Taiwan 13%, other AsiaPac 10%, Europe 10%, other Americas 2%
R&D expenditures in 2009:
$31.4 million
Market cap:
$207 million
≠ of employees:
200 including 106 in R&D and 60 in sales and marketing
≠ of customers:
1,000
Biggest customers:
Oracle/Sun the first one, followed by HP and Dell, then NEC, EMC/Iomega, Adaptec, Fusion-io, Seagate, Buffalo, Fabrik, LaCie, I-O Data, WD
Biggest distributors in 2009:
Excelpoint Systems (25% of sales in 2009), Promate Electronics(15%); Answer Technology and Avnet – biggest one in 2010 – (12%)
Main competitors:
ASMedia, IDT, Initio, JMicron, Marvell, and Symwave

                Key Financial Figures of PLX
                    
(FY ending December 31)

2009 Revenues by Products
         in Percentage

About President and CEO Ralph Schmitt

Ralph Schmitt, 41, was appointed president and CEO in November 2008. During 2008 he consulted with a variety of VCs, as well as acted as CEO of Legend Silicon, a privately-funded Chinese terrestrial digital TV semiconductor company. From 2005 through 2007, he was CE of Sipex, an analog semiconductor company, which merged with Exar in 2007 where he was appointed CEO. From 1999 to 2005, Schmitt was EVP of sales, marketing and business development for Cypress Semiconductor where he led sales growth to more than $1 billion and oversaw the acquisition and integration of numerous companies. Schmitt has also served on the boards at Cypress subsidiaries and other privately held semiconductor and systems companies. He began his career as a systems hardware designer and received his BSEE from Rutgers University. He is born 1960 in NJ, USA, and has two teenage daughters. Ralph Schmitt’s father died when he was young (a life-changing experience he cites as a source of his work ethic), after which he started work (at age 13) driving a forklift. He played university and professional soccer and was drafted by the North American Soccer League. He still plays in amateur leagues. Schmitt also plays keyboards for a rock-and-roll band, The Shanks, which won a Bay Area (California) radio station’s band competition and whose CD’s proceeds are donated to charity. He earned a Bachelor’s degree in electrical engineering from Rutgers University.

Q&As

StorageNewsletter: Since you are something of an interface specialist, how do you see the future of eSATA and FireWire 800 vs. USB 3.0?

Ralph Schmitt: You know, I believe eSATA will slowly fade away, because USB 3.0 performance will surpass eSATA, and it’s another connector that you don’t need on the PC. So we believe that interface will go away fairly rapidly. FireWire will probably be shortly thereafter. We’re going to support FireWire until it’s death, that’s one of the things we do well, we want to support that niche and we’ll continue to do that, but USB 3.0 will probably end up replacing that as well. Also, I think that after USB 3.0, there will probably be movement to more optical interfaces – Intel and Apple have talked about LightPeak, and we’re also looking in that direction as well. It would be one of the first optical interfaces in consumer products, so a key thing would be getting the price down for that.

USB 3.0 was supposed to be ten times faster than USB 2.0. Is that the reality?
I believe that it’s close to that. At least the testing that we’ve done. And it becomes more of a total system situation where the drives can start limiting the actual speed, so people will start optimizing certain drives to that interface, as long as the silicon can keep up, depending on the architecture that should be possible.

Yes, I’ve heard that some tests say just two times… but in the end, the problem is the disk drive, you think?
Yes, it can also be the software and the drivers. We’ve seen people have really bad performance based on that…

Currently, if you take just a 2.5-inch, 5,400rpm drive, does it saturate?
It should saturate, but it doesn’t always. You’d be surprised at the kind of performance you could have, because they down-bin also. But I’d say that performance will be limited not only by chip architecture, but by the hard drives, ultimately, and then software. That being said, we are seeing speeds up at the top level.

What’s going to change with PCIe 8Gb Gen 3?
So the bandwidth is very important in a few applications, the biggest one driving it is graphics, at 8Gb. So they need high throughput where the graphics processor guys are really starting to encroach into the data center more, with high intensity arithmetics – so they need a lot of throughput.

And compared to Ethernet…
Compared to Ethernet at 10Gb, with overhead, PCIe runs out about the same, because it has less overhead, roughly speaking its 70% data, when you look at Ethernet, 95% for PCIe.

What’s your roadmap in PCIe, NAS/DAS, and connectivity solutions?
PCIe is really Gen 3, plus some I/O virtualization features, we keep adding a lot of tools. In NAS/DAS, NAS is moving more and more to media, so we have more and more processors now, our solutions that we just brought to market, so we think that is going to give us higher performance and then adding applications, we’ve added things like PCI Express interfaces to the NAS products as well, so that gives us flexibility to do things. On DAS it’s USB 3.0, that’s really the future, we believe, and then ultimately into more optical solutions. And then connectivity, we’re combining things, so we’re doing a bridge there, and one of the interesting bridges we’re doing is PCIe to USB 3.0, so a lot of peripheral stuff, take a printer that has a PCIe bus, and they want to do USB 3.0 interconnect to it, so that’s where we see some interesting applications.

And about switches?
Yeah, well on PCIe’s switches, we see a trend into consumer applications, so small lane count, low cost. Because a lot of new applications need high-end processing, and so that’s a growth area for us as well. And printers are another example.

One final question – it’s not that easy for a company to work in a market that ranges from a small PC to the high-end, how can you manage that?
It’s difficult from a customer perspective, because there are lots of customers.

And even for your sales and marketing people…
Yeah, because it’s all around interconnect, but you get specialized in certain interconnects, and then you start combining, so that bridge I just told you about is a good place where you see there’s a combination of the two. Our biggest challenge as a company, and the reason we haven’t grown that much is because we haven’t been in big enough markets, so we’ve had to expand into multiple markets, but using the same technologies. And then we see storage as a good example, all the way from consumer to enterprise, we do the whole thing, like you said. But you have customers, like EMC/Iomega, and you’d be surprised by how much collaboration goes on between the two, and they want to drive the features set from the high-end NAS into the consumer NAS.