What are you looking for ?
Infinidat
Articles_top

Exclusive Interview With Steve Barber, CEO of Xyratex

"There was hundred of our systems underwater following Thai flood."

xyratex_barber Stephen (Steve) Barber, 51, is CEO of Xyratex since February 2003 and has served as a member of the board of directors since April 2002. He was previously president of the company from April 2002 and EVP and director of business development from October 2000. Prior that, he led the development of the firm’s warranty and repair business from a small UK-based company to a global business with operations in Europe, USA and Southeast Asia that was sold to Teleplan in June 1999. Barber acted as a division president of Teleplan from June 1999 before leaving Teleplan and re-joining Xyratex in October 2000. He was a member of the executive team that led Xyratex’ MBO from IBM in 1994 and previously held a number of management positions with IBM. He currently owns around 3% of Xyratex. He holds a Bachelor of Science degree in electronic engineering from the University of KwaZulu-Natal, Durban, South Africa. Hobbies: windsurfing, skiing and travelling.

StorageNewsletter: Your company has two completely different activities in storage: 1/ products for HDD process production, and 2/ disk controllers and storage subsystems, without any relation in term of market and customers. Why did you never separate these two businesses into two different companies?
Steve Barber: If we look back at our history with our heritage in disk drives manufacturing and system development, we saw the opportunity to address two parts of the storage marketplace, Very different technologies but with the common ingredient of being the disk drives and, with our core expertise of understanding how to utilize them, to create a test environment and equally integrate them in an enterprise class storage system. There is a lot of common “know how” with regard of heat management, vibration management, interface, protocol, etc.

And you have no intention to separate these businesses ?
We’ve clearly had that discussion at a board level. Our general view is that Xyratex is not a large cap company. We’re a public entity and separating the two businesses would make two even smaller businesses as seen by investors. Therefore we maintain the way it is.

When you were manufacturing IBM HDDs. how many makers were there at this time?
There were a lot, lot of companies who have been absorbed, like Maxtor, there probably was dozens.

Are you happy now that there are only three?
I think it makes for a less volatile marketplace. From a Xyratex standpoint, we have good relationships with all three of them, and I think, as you’re seeing right now the HDD marketplace is more stable, supply management is much more stable. I think we have a much more mature market and that is good for us and the industry. We are all very reliant on storage, we need a good, solid, disk drive industry to support all other IT. I think it is an inevitable outcome of where we’re going, the cost of development, the cost of new technology demands less players

You work apparently with Seagate, Toshiba and WD for HDD equipment, but will you get Toshiba for 2.5-inch HDDs?
We have Toshiba today for 3.5-inch in enterprise, we would like to be seen as a good suppliers for 2.5-inch.

The global production of HDDs continue to increase with less makers. But do you sell more testers?
The demand for test equipments is driven by two factors, quantity of HDDs shipped and the capacity of the HDDs shipped. Provided we supply to all companies then we’re protected from whoever has the biggest market share. Right now and for the first time in history, we’ve seen this year the volume of HDDs declining because the sell of laptops is slowing down, therefore the demand for 2.5-inch is going down, however, the capacity of HDD is increasing significantly and as a result the total test capacity required continues to increase.

Last quarter your revenues were up to 85% compared to the prior quarter. Are you also one of these companies benefiting from the dramatic Thai flood?
Yes, 2011, 2012 was a unique period for the HDD industry. We saw two major events in 2011 that impacted these sale of capital equipment. First of all, the Japanese tsunami followed by the Thailand flood, which were more devastating to the industry. And as a result the industry wasn’t able to produce HDDs, probably about 1/3 of HDD capacity was removed with the flood. And for Xyratex it was a very difficult year, very limited demand for production equipments. As you’ve seen this year we were actively involved in the struggle to assist them in repairing, replacing in some cases equipments that were damaged in the flood.

Do you know how many of your equipments were in the flood?
There was approximately 100 systems underwater. WD before the flood took defensive actions to remove as much of the components of our systems out of them and take them to higher ground before the damages. So we were able to re-use a lot of the materials on new systems. But equally there was some systems that were damaged. So we’ve been working with them to re-utilize previous equipments but also to supply new equipments for them.

Will you enter into SSD testers?
I think that’s unlikely. Our expertise is in managing the unique characteristics of HDDs, thermal issues, vibration chambers. SSDs is a piece of silicon with an interface and that doesn’t really play to our expertise. We recognize SSDs are becoming an important part of the storage portfolio. We don’t envisage, certainly in the next 5 to 10 years, that SSDs will replace HDDs for the vast majority of storage. Today if you look at the production capacity for SSDs in terms of fab, they contribute approximately from 3% to % of the worldwide petabytes of storage requirements. Even if they are free, they would only be able to provide three percent of the need for storage. If storage is growing at 40% compound in the growth-rate, I don’t see fab capacity growing at that rate. The cost of investing in fab is phenomena, 10 billion dollar, would only double their forces, maybe they would be able to fulfill 5% or 6% of the requirements.

You claim that about 50% of the WW HDDs are produced with your technology. But who are your few competitors?
Combination of internal design by HDD companies themselves. Hitachi for example, who uses subsidiaries of Hitachi to produce equipments, that’s changing. But they have a huge installed base of  their own equipments. We have seen others competitors in the past, company likes EOtech providing equipments to OEM. More recently we have seen competition with Teradyne, they entered the space about 2 years ago and today they are producing 2.5-inch HDD equipments that we compete directly against.

According to IDC, Xyratex is the largest OEM disk storage system provider? Can you prove that?
All IDC does is adding up the OEM sales revenue for all of the OEM businesses. Include in there is Dell OEM EMC products, the CLARiiON product line. That product category was classified as OEM. EMC was an OEM supplier.We compete against those companies. When this report was first compiled back in 2008, I think we were at 50% market share, in the 2010 financial we were almost at a 50% market share. In our 2011 results, issued in November or December I think we will be over 50% in OEM sales. Probably ahead of LSI Ingenio, EMC, HDS. There are resellers agreements not OEM agreements. Today as you know, NetApp OEM product is IBM, IBM NAS products are actually NetApp boxes, and actually they are Xyratex OEM products. But that market on OEM is only about 20% of the general market. 80% of the market is controlled by in-house manufacturing.

What do you do for each OEM, for which products?
The FAS3000 product range and the FAS6000 are our main products for NetApp. We work with Dell with the EqualLogic and Compellent product range. Our OEM IBM products are the V7000 and XIV. And we did business with EMC for approximately two years doing all of their Data Domain products. Around May of this year that product ended and EMC moved that software onto an EMC product. Therefore EMC is no longer on board with Xyratex platform.

You seem to sell a software called OneStor. Do you intend to be also a software storage company? For which application?
OneStor is a hardware platform not a software platform.We don’t sell HPC software alone, we sell a software management system along with Lustre integrated into the ClusterStor platform.

Today, are you more of a hardware or software company?
We are increasingly a software company. We choose to use software to produce solutions as opposed to selling software as a standalone product.

You entered recently in storage for HPC following the acquisition of ClusterStor in 2010. Does it mean that, for the first time, you will sell to  non OEMs and directly to end users?
Our root to market for HPC solutions has been to co-sell along with our OEM partners.

So always through an other company?
Correct. Today you’re aware of our relationship with Cray and we’re entering into similar agreements with additional companies. We haven’t publicly announced these companies. We have partnered with Bull and certainly value them as a partner and equally there are other major computer brands that we will partner with. Potentially as an OEM provider, in some cases, depending on the needs of these companies maybe as resellers so those company would resell the Xyratex products.

Your main competitors in HPC storage ?
Would be DataDirect Networks, followed by Panasas.

Historically you’re specialized in interfaces, will you invest in Thunderbolt?
We don’t see that as a core business going forward. No.

What’s your roadmap?
Today, because of the breadth of our business, our roadmap is addressing a number of different parts of the storage marketplace. So we have a very clear roadmap in the capital equipment business which is all about increasing productivity benefits for our customers, improving footprints, power consumption, space utilization. We’re looking to our OEMs business. Huge focus on power supply and addressing a range of emerging companies, partner with companies like StorSimple or Pure Storage providing a range of storage devices but increasingly with an embedded CPU. We have a range of what we call "Application Platforms" with a roadmap of low-performance, high-performance really far in the Intel evolution of chips. Increasing power, increasing speed and we’re looking in the HPC space and we see that market extending in clearly higher performance products. But beyond that we are now aggressively looking to address the cloud infrastructure marketplace. Very high capacity, very high dense rack solutions.

What’s the next technology you will invest in?
We’re moving toward HAMR recording. HAMR is extremely complex and it’s one of the reasons why companies have consolidated the investment cost. We are working very closely with these companies today to assist them in developing HAMR production and test capabilities, but that technology is extremely challenging, not just the technology but also the cost. It needs to be extremely cost-efficient. We are also focused on 12Gb/s SAS interconnect as well as cloud storage infrastructure.

Since the first quarter of 2011, your global revenues are decreasing. What’s the problem? When will you be back to growth?
In there is very much the feature of 2009 recession, created a huge catch-up in the marketplace in the early month of 2010. We saw more revenue and demand for store system and capital equipment in the early part of 2010 which was not normal, it was a catch-up after all of the budget cuts and the constraints of the 2009 recession. Moving to 2011 we’re faced with the tsunami and the flood which had a very significant impacts to our capital equipment business compare to 2010. In 2012 we’re seeing a recovery in capital equipment as we’re now replacing and repairing damages. I think we will start to regain power and momentum late 2013, start of 2014.

For next quarter your prevision is between between $313 and $373 million. It’s really broad, wide. Does it mean that you don’t have a clear vision of your short-term future?
All of the major customers are cautious and whilst we have clear certainties, real purchase holders, commitment of delivery date there are still variables and in this current macro-economic centric environment, consumers spending decline, we felt, based on feedback of our customers that providing a wider than normal range was the right thing to do. It is changing. We see some customers becoming more confident and we see others backing off. It’s a very uncertain market.

What’s your reaction if I tell you that Xyratex is a poor marketing company?
I would probably agree with you. We have not worked to raise our brand awareness. We have been very careful to raise the profile of our customers names and brands, not Xyratex. That’s very conscious, we are OEM supplier. We are recognized in the industry and people who are entering the storage space will know Xyratex.

Do you have an acquisition strategy considering that you didn’t buy any company since December 2010?
We continue to be opportunistic in software. We see a number of areas as we look toward the cloud business where there are software requirements. And today we do not have a full portfolio for these situations. We have a good history of acquiring IP-based companies not revenue streams. Our focus is smaller company, more in buying people expertise.

(Transcription by Corentin Béchade)

Documents on Xyratex

Milestones

  • 1966: Beginning of business under IBM manufacturing HDDs
  • 1980: Enters in storage systems
  • 1994: Becomes independent following MBO
  • 2002: Incorporated under the laws of Bermuda
  • 2004: IPO on Nasdaq, raising around $50 million
  • 2008: Revenues exceed $1 billion
  • 2011: Enters in storage for HPC

                                 
Acquired companies

Year  Acquisitions Price* Activity
1995  Peer Protocols   NA Test equipment for SCSI protocol
1996  White Electronics   NA Dutch warranty and repair of monitors
1996  BigByte Europe   NA Floppy disk and warranty management
1996  Zadian
 Technologies
  NA Broadcast technology
and design services
2004  ZT Automation   NA Handling automation systems
for hard disks
2004  Beyond3
 1.3+ Optical inspection system
for HD manufacturing
2005  NetHelix (assets)
  NA HDD component test
2005  Oliver Design  17.2 HD media cleaning product
2005  nStor  21.2 Entry-level RAID controller
2006  Jastam Trading   1.5 Distributor in Japan
2006  Ario Data
 Networks
  NA Bladed storage solutions
and RAID controllers
2010  ClusterStor   NA Clustered and parallel
storage system for HPC
2010  Optical Systems
 Corp
   5 Automated production technology
2010  Magnetic
 Recording
 Solutions
  NA HD media and head testing

* in $ million

                          
Activities sold

Year  Acquirer Price Activity sold
1999  Teleplan   NA Warranty and repair
2004  GlassHouse
 Technologies
  NA Sagitta Performance Systems,
in storage networking products
and storage consulting
2006  Napatech   NA 1GbE X-cards for network analysis

                                 
Yearly Revenues
xyratex_revenues_540

Employee base


As of end of June the company had around 2,000 permanent employees and 1,000 contractors, located in the following regions:

  • Havant, Hampshire, UK (HQs) – 700, R&D, manufacturing, sales, purchasing
  • USA (CA and FL) – 750, R&D, manufacturing, sales
  • Seremban, Malaysia – 700 (plus up to 700 additional contract staff depending on workload activity), manufacturing, some R&D, Asia region HQ
  • Mexico – 35, manufacturing
  • Canada, India, Romania, Russia and Ukraine: Software development employees
  • China, Malaysia, Philippines, Singapore and Thailand: HDD capital equipment support teams

Articles_bottom
AIC
ATTO
OPEN-E