Rackspace revenue continued to rise during the third quarter, but growth was slow and profits were down year over year. The company chalks up the latter to increased forward-looking investments, but the elephant in the room is Amazon.
Software-defined networking vendors such as Embrane and Nicira have found customers in the managed-hosting realm, and with more startups bringing products to market, enterprises could follow suit later this year.
Rackspace has added virtual desktops to its collection of cloud services, a move that could prove very lucrative as the worlds of cloud computing and next-generation mobile devices converge. The new offering, called Hosted Virtual Desktop, pairs Citrix’s XenApp and XenDesktop products with Rackspace infrastructure.
Rackspace has made a name for itself as a provider of managed hosting services, by being an increasingly strong competitor in the public cloud market and as a key player in the open-source OpenStack project. This week, the company announced its new Cloud Builders business, in which a team can enter other companies’ data centers and install OpenStack for them. But with this move, has the company strategically broadened its portfolio or damagingly muddled its brand?
Established in 1998 as a hosting company, Rackspace entered the cloud storage market with Mosso in 2006. Mosso rebranded as Rackspace Cloud in 2009, and its technology became the current Cloud Files storage solution offered today. The Rackspace Cloud also includes a processing product known as Cloud Servers, based on technology Rackspace acquired with Slicehost in 2008. Figures collected by Onavo CEO Guy Rosen suggest that Rackspace is number two in the cloud-server business, behind Amazon. Those figures also imply that this gap may be closing.
But despite growing revenues, it’s unlikely Rackspace will pass Amazon soon, and perhaps recognizing that, the company joined with NASA and others to establish the open-source OpenStack project. The community behind the project has grown, and now includes over fifty hardware, software and solutions providers, including Canonical, Cisco, Dell and Citrix. Rackspace donated its Cloud Files code to the project, which became OpenStack Object Storage. NASA contributed code — largely sourced from its contractor, Anso — from its Nebula Cloud, which became OpenStack Compute. Together, these products offer an increasingly robust open-source solution for those wanting to build and run clouds.
Lending corporate weight to OpenStack makes sense for Rackspace on a number of fronts:
- The project sits logically alongside the company’s existing hosting and cloud businesses.
- The donated Rackspace Cloud Files code is improved and extended through the efforts of Rackspace’s many partners in OpenStack.
- These enhancements are incorporated back into the commercial Cloud Files product, making it better, too.
- Developers and users gain familiarity with the Rackspace Cloud Files/OpenStack Object Storage methodology, which gains traction as it is used by Rackspace and every non-Rackspace site to which OpenStack is deployed.
- Rackspace’s commercial hosting and cloud products become the logical solution for those seeking support and other features not available from a smaller open-source deployment of the code, increasing market share and revenue.
Each of these steps is logical and defensible, and complement the existing products in Rackspace’s portfolio. Despite the company’s acknowledged reputation for “Fanatical Support,” this week’s announcement of Cloud Builders strikes me as a step that is far less clear than those that went before. It is worth noting that my colleague Derrick Harris seems to disagree, writing on Tuesday that “today’s news should be good news both for OpenStack adoption and for Rackspace’s bottom line.”
With Cloud Builders, Rackspace offers three services:
- Support for those trying to deploy OpenStack on their own hardware, which fits well beside Rackspace’s existing support team.
- Training and certification for those wishing to design, deploy and maintain OpenStack clouds; it may be possible to align this with Rackspace’s existing support apparatus.
- Deployment services that will go out to customer sites, design a cloud infrastructure and then plug the requisite hardware and software together to get a new cloud up and running.
It is this final service that competes with the company’s existing businesses, possibly drawing customers away from Rackspace’s established hosting and cloud products. It brings Rackspace inside the data center, into direct competition with solutions providers previously might have recommended Rackspace’s off-site products; those competitors could retaliate in the future by recommending other hosting or cloud providers. Finally, and far more seriously, the announcement about Cloud Builders creates confusion in the market about Rackspace’s intentions. It’s a model that is expensive, labor-intensive and difficult to scale — very different from Rackspace which today operates at scale and with low margins.
Cloud Builders only subtly changes Rackspace today, but it also leaves me uncertain about what the company wants to be tomorrow.
Question of the week
Verizon announcing plans to buy Terremark for $1.4 billion is a big deal in the cloud computing world, but I don’t think Terremark will be the biggest winner as a result of the acquisition. Sure, Terremark gives Verizon an even larger global footprint, hundreds of millions in annual revenue and a relatively innovative cloud business that makes Verizon look a lot better against telco competitors such as AT&T, but the key word is relatively. There’s still a big difference between cloud innovation at the MSP level and that at the pure cloud level, and this deal just widens the gap between types of cloud providers. Both have their places and, now more than ever, Rackspace looks great as the company straddling the line between these two worlds.
Here’s what I mean. MSPs offering cloud services (e.g., Terremark, Savvis, SoftLayer, Peer1, Internap, etc.) offer valuable services for large businesses, but nobody is mistaking them for cloud-first, or pure-cloud, providers, pioneers, such as Amazon Web Services, Microsoft, GoGrid and Joyent. I wrote recently that Gartner got it wrong with its Magic Quadrant for Web Hosting and Cloud Infrastructure as a Service because the two really are two different beasts, and I stand by that position. MSPs have a variety of hosting options both dedicated and shared, enterprise-grade contracts, skin-in-the-game SLAs and general support that make them very appealing to large companies and governments that want to move workloads to the cloud, but their primary businesses are not cloud computing.
Terremark is a prime example of this. Its Enterprise Cloud offering has a number of notable customers, including from the U.S. government, but its vCloud-based self-service cloud is somewhat of a red-headed stepchild. From what I’ve heard, Terremark’s cloud business comes mostly from, and its energy goes mostly toward, the Enterprise Cloud, which relies on many of the enterprise-standard practices mentioned above. Overall, Terremark is predicting cloud revenues of $30 million for its fiscal year 2010 (which ended Dec. 31) out of total revenues predicted to be $350-$353 million. At best, that’s 8.5 percent of total revenue. Verizon buying Terremark is a case of a more-conservative company buying a pretty conservative company.
Cloud-first providers, on the other hand, generally don’t have dedicated infrastructure (one notable exception is GoGrid, which actually rents dedicated infrastructure just like shared cloud infrastructure) and focus their operations on providing a true self-service experience far removed from the traditional IT procurement cycle. Whatever they’re offering, it’s a cloud-first message, and it’s all cloud revenue, so all energy and innovative spirit are put toward cloud computing. This approach is more appealing to developers, and even smaller companies without strict regulatory compliance requirements, who want real change in how they buy and procure IT. Granted, it has a larger percentage of a smaller pie, but it’s estimated that Amazon Web Services revenues topped $500 million in 2010 – more than the entirety of Terremark’s revenue.
Right smack in the middle of both these camps lies Rackspace, which amassed $565.8 million in revenue during the first three quarters of 2010, of which 12.2 percent – $69.2 million – came from cloud computing. The latter number has been growing with each quarter and should only continue to rise as Rackspace furthers the integration between its stalwart managed hosting business and its cloud business, and when its much-ballyhooed OpenStack cloud-platform software becomes production-ready. At that point, many think Rackspace will be the largest in an ecosystem of cloud providers all running the same foundational software, which has developers very excited. One recent analysis already has Rackspace neck and neck with Amazon Web Services in terms of the number of top 500,000 web sites hosted.
Rackspace was among a number of MSPs whose stocks climbed on Friday after the Verizon-Terremark deal was announced on Thursday, and for good reason. Terremark is a relatively small player among MSPs. By contrast, Rackspace’s market cap is already at $4 billion, and it’s only going to keep earning more money from across its business lines because it caters to everyone from developers to large businesses. Competitors such as Savvis – another company possibly for sale now – presently do more total revenue than Rackspace – Savvis did $680.3 in total revenue during the first three quarters of 2010 – but managed hosting accounted for only 46 percent of third-quarter revenues, and cloud computing is just a part of that number. If Terremark ends being just the first domino to fall, and if cloud computing is the driver, Rackspace might be the biggest.
Related Research: Why New IaaS Providers Enter at Their Own Risk
Question of the week
Cloud provider GoGrid has expanded its Infrastructure-as-a-Service catalog by launching a Hosted Private Cloud that maintains all the features multitenant clouds, but on dedicated physical servers. It’s an interesting tactic, and it highlights the different value propositions and visions of the leading cloud providers.
I gotta say, I was pretty surprised this morning to read about NephoScale, a new cloud provider combining IaaS through both virtual and dedicated servers. The IaaS landscape is crowded as is, and I don’t know what a new provider can really to to distinguish itself. From AWS, sure. But from AWS, Rackspace, Windows Azure. GoGrid, OpSource, SoftLayer, Savvis …? I wish NephoScale luck, but if wants to be more than a niche or regional provider, it needs to keep the new features coming fast and furiously.
Amazon Web Services has made available two additional support options for customers of its cloud computing services. Customers can now choose from the Bronze level, which costs $49 a month, or the Platinum level, which costs at least $15,000 a month.
Yesterday on the Structure blog, I ripped Gartner’s Magic Quadrant for Cloud Infrastructure as a Service and Web Hosting for not including Amazon Web Services among the leaders. Last night, Gartner analyst Lydia Leong, co-author of the report, responded and defended her rankings. I think she makes some fair points about what Gartner clients want and why AWS isn’t necessarily enterprise-friendly, but I still think the report is misleading. If we’re ranking hosting providers that have cloud offerings, then AWS shouldn’t be included at all. But when ranking IaaS, AWS has to be at the top. It’s the reason all other providers have cloud offerings at all, and its pricing and self-service models — which continue to evolve — have revolutionized our conceptions of IT, even if all users haven’t yet caught up.
Rackspace’s new AppMatcher marketplace is intriguing for a variety of reasons. For one, it smacks of the Microsoft Windows Azure strategy of bringing together users and ISVs under a single roof. The strategy is that a big enough user population means everyone will want to participate. And while you’re experiencing the Rackspace goodness, why not host your application there too, or maybe offload other workloads like you are the one being delivered via SaaS. Where Rackspace misses out, though, is in guaranteeing developer dollars — if it had its own platform, apps would have to reside there. As is, it must hope it can lure ISVs into becoming paying customers.