DCD FOCUS: You have an interesting background. You began your working life selling organic pork, then moved into real estate. How did you take the leap into data centers? What was your vision at the time?

Rackspace co-founder and non-executive chairman of the board Graham Weston: I was fortunate to grow up in a family of entrepreneurs: my dad, my uncles and others. The organic pork business was one that I launched as a kid, when I was active in Future Farmers of America, an organization that’s popular among farm and ranch kids in the US. What gave me the idea was the big concern at the time over chemicals being added to pork products. I sold plenty of organic pork in my little venture but later found that I didn’t charge enough to cover my expenses. That’s a lesson I was glad to learn early in life. 

Right now, the big commodity cloud providers are chasing each other to the bottom on pricing. But at Rackspace, we sell a premium product – managed cloud. We’re targeting business customers that value managed services and our Fanatical Support, and who are glad to pay for the value that those services add.

During and after college I got involved in various businesses and investments related to real estate in the San Antonio area. But I was always interested in technology.

The idea for Rackspace came from three young men who had been students together at Trinity University in San Antonio. When they first pitched their idea to me, as a potential investor, I saw it in real estate terms. We would be renting out computer hardware the way you would rent out apartments or office space — except that we would never run out of inventory. We would just buy more servers as we attracted more customers. Our tenants would be businesses that wanted to rent access to computing and storage resources, instead of making big capital investments in hardware and software.

Our vision at the beginning was very different from the one we have today. We had what we called a ‘denial-of-service’ business model. Right up front on our website we warned potential customers that they had better be technically capable enough to work from a command line and handle their own IT problems because we weren’t going to help them. In real estate terms, our vision was that we would just collect the rent and the tenants would unclog their own toilets.

What we quickly learned was that even highly technical customers needed help with computing. The technology was complex and fast-changing. At the same time, we were getting more competition from hosting companies that didn’t provide any customer service. We needed to differentiate ourselves. So we decided to change our whole vision of the business, to provide what we called ‘Fanatical Support’. We started answering the phone and helping customers to succeed. We started recruiting geeks with a great bedside manner. We expanded our areas of expertise in Linux and Windows and areas like network security. We became profitable early on, and grew rapidly. 

Operating a data center must have brought some very different real estate challenges and pushed you further into the tech arena then you had ever been before?
In the beginning we operated out of one floor of a high-rise office building that my family owns in downtown San Antonio, just a few blocks from the Alamo. From the time we opened our website for business, when we had just a handful of Rackers (as we call our employees), we have operated the business 24/7. For a while, we had the night shift folks resting in sleeping bags on the floor near the servers. As we grew, we began to move into data centers around the country, and eventually around the world.

We were fortunate in a way that we were growing right after the dot-com bubble had burst, so we were able to buy and lease data centers at low prices. It was fascinating to me to learn the economics of the data center side of the business, which fit well with my background in real estate. I left the technical side of things to others.

You asked about challenges, and one the biggest that we ever faced was one night in 2009 when a passing truck crashed into a utility pole and transformer and knocked out power at our main data center in the Dallas-Fort Worth area. We had backup generators but some of our chillers didn’t come back up. We had to shut down most of our customers for an extended period. That tested our commitment to Fanatical Support and transparency with customers. As more and more Rackers heard about the outage they rushed into the office without asking anybody, many with their kids in tow. They stayed into the night, answering phones, reaching out to customers, eating cold pizza and helping to get the data center back online.

That experience made us determined to make the quality of our infrastructure second to none. We added a lot of safeguards like redundant network interface cards, multiple redundant power supplies, redundant switches in the cabinets and hot-swappable components. Our servers come standard with local RAID10 SSDs, which increase your performance and reduce the likelihood of failure. And we use top-of-the-line Error Checking and Correction (ECC) RAM – the type used when data corruption is unacceptable.

These safeguards have helped us achieve 99.999% data center uptime for the past several years. These are safeguards that few of our competitors have put in place. But we have a different philosophy than they do about redundancy.

Any cloud provider, Rackspace included, will tell you that instances in a multi-tenant cloud should be thought of as easily replaceable, and that you should architect your app to protect it against hardware failure. Some commodity cloud providers go further and regard infrastructure quality as not very important. If a commodity provider has an outage in a data center, or part of one, and the customer goes down, the provider considers that the customer’s fault. They say that the customer must not have ‘engineered for failure’ across multiple availability zones and DCs. (And never mind that many customers don’t do so, usually for cost reasons.) The provider’s SLA often makes clear that such a customer can expect no compensation for his downtime.

At Rackspace, we take a different approach. We know from long experience with customers that a failed instance can be disruptive, whether you run a solution architected across multiple data centers or run a single server in a single data center. We want those failures to happen as infrequently as possible. We see redundancy as a shared responsibility. We guarantee uptime and we back it with industry-leading SLAs. If you go down, it’s as much our problem as it is yours.

Today Rackspace finds itself part of a rapidly growing industry in which it already has a strong foothold. So why is the company now exploring options for partnerships or acquisition?
Regarding what we are doing in terms of exploring options for partnerships or acquisitions, I can only direct you to our Form 8-K, which we filed with the Securities and Exchange Commission on May 15, 2014. (The form, released in May, regards Rackspace’s hiring of Morgan Stanley to evaluate strategic partnerships.)

I will generally state that it is always the responsibility of directors and officers of a public company to assess potential business opportunities to determine whether they believe the opportunity will help advance the mission. We believe in our mission here and have always carefully assessed opportunities as they have arisen and will continue to do so.

I would add that we are very confident in our prospects and in our positioning. We’ve made clear that we’re playing a different game than the big commodity cloud providers. They’re renting out access to raw infrastructure and expecting the customer to do all the engineering work to run that infrastructure, not to mention all of the complex tools and applications that run on top of it – for everything from data engines to ecommerce platforms.

We’re taking a very different approach. We’re helping customers that want the power of the Cloud without the pain of becoming expert in dozens of cloud technologies. We’re focused on customers that want to stay fast and lean – to focus their precious engineering resources on their core business without swelling their payrolls with engineers to manage cloud operations that don’t differentiate their company. We give those customers managed services that they can’t get anywhere else, along with specialized expertise and support.

Do you foresee that size – or footprint – will offer the key competitive advantage in future?
Size will continue to be important in one part of the cloud computing market – the part served by the big commodity providers like Amazon and Google and Microsoft. Those are great engineering companies. They offer prodigious economies of scale and they’ve done a great job of lowering the cost of raw infrastructure for customers willing and able to do all the engineering work to operate that infrastructure and all the complex apps that run on top of it.

We operate in a very different segment of the market. We offer something different, which we call economies of expertise. We help customers operate their cloud infrastructure so they can focus on their core business. We also offer them specialized expertise in complex technologies where that expertise is expensive and hard to come by — for example, in data engines like MongoDB and Hadoop and Redis; in ecommerce platforms like Magento and Hybris and OracleATG; in digital marketing applications like Adobe Experience Manager and Sitecore; in collaboration software like Sharepoint, and in running private clouds based on OpenStack, the open-source operating system that we launched in partnership with NASA.

We’ve built up expertise and credentials in each of these areas. We also run the world’s largest OpenStack public cloud and we’re “leading the pack” in running DevOps automation tools for our customers, according to Gartner analyst Lydia Leong.

You ask a good question about footprint as an advantage. We do think that’s an advantage in serving certain customers for whom low latency is a big priority. That’s one reason we’ve expanded our footprint with multiple data centers in the US and the UK, and with others in Hong Kong and Sydney.  We’re always examining options for further expansion to serve continental Europe and other parts of Asia as well as Latin America.

We also differentiate ourselves from Google and Amazon in two other ways. First, we’re committed to open and standard technologies, so that our customers never feel locked in to one provider, the way they are on big commodity cloud providers that use proprietary technologies. That’s why we partnered with NASA to launch OpenStack, the open-source operating system for cloud computing. That’s why we’ve optimized our cloud to run open and standard technologies.

Second, we offer hybrid cloud computing that gives each customer the best fit for its unique IT needs.

Our broad product portfolio and our specialized expertise help each customer run each of its workloads where it will perform best and most cost-efficiently, whether that’s on single-tenant or multi-tenant servers, or on a combination of those platforms.

We’re glad to serve customers in data centers owned by us, by them, or by a third party. Our hybrid cloud approach is very different from the one-size-fits-all approach taken by the big commodity cloud providers. They only offer multi-tenant public cloud. They argue, against all the evidence, that every workload runs best there.

Even though we play a different game than the big commodity cloud providers we do compete with them for a certain type of customer. That customer typically got going as a startup on, say, AWS, because it was cheap and easy. Then that customer starts growing rapidly and getting big. And they find that the one-size-fits-all approach starts binding and chafing around the crotch and around the wallet pocket. They find that their ‘savings’ on low unit prices for compute are overwhelmed by bigger costs. They are constantly in contention for resources in the multi-tenant public cloud. They’re competing with noisy neighbours on the disk and on the network. And they’re paying a hypervisor tax that affects their performance. To boost their performance they end up overprovisioning, which leads to nasty surprises in the monthly bill. Many of them consider moving to a colocation facility for more predictable and consistent performance and cost, as well as for advantages in control and security and compliance. But they’d rather not give up the agility and flexibility of the public cloud.

For these customers we recently launched a new product called OnMetal, which is revolutionizing the way that production applications are run at scale. It lets customers spin up – and spin down – single-tenant, bare-metal servers as quickly and easily as they can spin up virtual machines. They can access those bare-metal servers through the same OpenStack API they use on our multi-tenant public cloud and they pay only for the computing they use, just like in the multi-tenant public cloud.

They can get the best of both worlds and they can do so on an open technology.

OnMetal is built on OpenStack software and OpenCompute hardware, which will allow for rapid improvements to be made by our collaborators in those open-source communities. OnMetal has been available in general availability in the US from July 24. This gives customers the chance to test and plan how to take advantage of this exciting technology before it is available in 2015 in Rackspace’s international data centers.

Will Rackspace be looking to make acquisitions of its own around OpenStack?
We are always examining potential acquisition targets, mostly with an eye to acquiring technical and leadership talent and new technologies that will expand our specialized expertise in areas customers ask for help. Most acquisitions to date have been relatively small but we think they’ve given us good bang for the buck.

Our acquisition of ObjectRocket, for example, gave us invaluable expertise in MongoDB and other data engines, and a respected leadership position in the technical communities that advance those technologies

Our acquisition of Sharepoint911 did the same in hosted Sharepoint.

We tend not to make acquisitions just to get bigger or to add revenue.

Now, getting back to your real estate roots, what do you think will be Rackspace’s real estate challenges in the years to come? And what is driving these?
I think we’re in good shape on the real-estate front. We continue to grow and hire at a brisk pace.

We’ve got good plans in place to accommodate that expansion at our headquarters in a renovated shopping mall in San Antonio, and at our offices around the world. We recently announced that we’re in negotiations to move our Austin office into – you guessed it – a renovated shopping mall.

We’re big believers in urban infill projects, and in doing what we can to enrich the communities where we do business.

How do you see the overall market evolving to keep up with demand? Do you see more consolidation of competitors, changes in technology to allow for more hybrid environments and other technology areas leading to sweeping changes in the hosting/cloud business space?
When any market is young and fast-growing it is dominated by enthusiasts and early adopters, and the providers that cater to them.

Then as the product or technology catches on, it brings a second and larger wave of adoption from mainstream customers.

Then the market starts to segment, with different providers serving customers with different needs and preferences.

You can look back to the early automobile industry as an example. The early cars were not very reliable. It was a pretty difficult process just to get them started. There wasn’t much of a service infrastructure, so early adopters had to learn how to maintain and repair the cars themselves. The service manuals ran to hundreds of pages.

That was fine with the early adopters and enthusiasts. But there was a larger potential market that just wanted to use a car, without having to become an expert in how to take it apart and put it back together again.

Before long, the car makers started appealing to that second wave of customers. They made their cars more reliable and easier to use. Dealers and gas stations added more service technicians. And eventually you had not just gas-powered cars but diesel trucks and hybrid vehicles and electric ones.

In cloud computing I think we’re just starting the second wave of adoption.

Until now adoption has been driven by tech enthusiasts, many of which like to tinker with their own cars – if you will. They like to do everything themselves, even if that distracts them from working on their core business – on their application. But that’s changing rapidly.

Right now, the best estimates are that about 80% of all computing is still done on hardware owned and managed by the users. Most of that computing is moving toward the Cloud, but not necessarily directly to the multi-tenant public cloud. Many of the pragmatic businesses in that second wave will want a hybrid solution that might involve some multi-tenant public cloud, some single-tenant servers and maybe some hosted VMware.

Those companies want to focus on their core business. They want a trusted partner to manage their cloud.

Now to be sure, there will continue to be customers that want to rent access to raw infrastructure and do all the engineering work required to run that infrastructure themselves. Amazon and Google and Microsoft will continue to thrive.

But I think there will be more and more customers that want a managed cloud.

I feel very good about where we are positioned at Rackspace, and about the specialized expertise that we’re developing in the areas where customers want help.

I think we’re well-positioned right smack in front of a big new wave of managed cloud adoption that will last for many years. I think we’ll be able to serve those customers well, and add tons of value and help them succeed

That’s what makes me jump out of bed every morning and show up at work with a big smile.

This article first appeared in Focus 37. To read the digital edition click here.
Since the article was published Rackspace has declared its commitment to remain independent after reports it was looking for a bigger partner to acquire it. Rackspace has also appointed a new CEO, Taylor Rhodes, to lead and drive its managed cloud strategy.