This journey is something service providers and telcos struggle with as well – they know the power of the cloud, their customers are asking for it, but they’re not sure where to start. So many questions… Should they pick a hardware vendor and buy a truck full of gear? A software vendor and buy some licenses? Some kind of appliance? An integrator to custom-build them something or to plug those things together? How will they package, productize, price, sell and support these offerings? How will they build an engineering and operations team to maintain the high levels of SLAs their customers will demand?

Lots of questions, no great answers. Until now!

Today, Rackspace is telling the world that we plan to build and run a linked global cloud network for large service providers and telcos. This cloud network will run the Rackspace public cloud – the largest OpenStack-based public cloud in the world – at global scale, creating a networked cloud of clouds through which service providers are linked together behind the scenes.

Rackspace will deploy our public cloud into a service provider data center and remotely operate that cloud, and the service provider will provide the physical data center operations as well as market, sell and support the co-branded cloud directly to its end-customers. The end result is more OpenStack in the hands of more users, which has been the Rackspace vision since we founded the project in 2010.

This extension of our public cloud portfolio will give both service providers and their customers access to a network of interconnected global data centers – customers will be able to run their workloads in any data center that is part of the network, while maintaining their business relationship with the local provider of their choice. It will also expand Rackspace’s global data center footprint to include these service provider partners, ultimately letting us and our partners increase resource utilization and cloud growth. Users will be able to run their workloads where they want, when they want.

As operators of the world’s largest OpenStack-based public cloud, we’ve gotten very good at running clouds at scale. We’ll share that expertise with our service provider and telco partners, giving them everything they need to build an OpenStack cloud offering: we’ll design it, deploy it and operate it. They’ll have access to our expertise in technology, operations, go-to-market and support. It will be powered by OpenStack, and we’ll provide the patching, tuning and monitoring backed by carrier-grade SLAs. We’ll give our partners sales and support training, collateral and more to help them successfully enter the cloud market faster while helping to reduce their business risk.

Now, service providers and telcos can broaden their portfolios to offer their customers cloud services on a global level. They can complement their existing services with a proven public cloud offering that is powered by open source software and taps into the vast expertise of Rackspace.

This is our vision for the cloud – truly interoperable open clouds at global scale. We’ve been asked for it by service providers on nearly every continent. And now we have a team working hard to deliver it.

This is a major step in helping OpenStack gain wider adoption – and we’re building it to better serve the world and our customers.

Questions or comments about this blog post? Let me know below or on Twitter: @scottsanchez

Two years ago, when we first started in the private cloud world supporting clouds in customers’ data centers, we operated as a consultancy.

We took raw OpenStack code and used our expertise in running a large public cloud to build custom private clouds for customers.

Every customer had a different idea of what their private cloud should look like, and we’d end up helping them build what I like to call “Crazy Cloud™” (trademark inflection is my own). You can define Crazy Cloud as an unsupportable mixed bag of bits that fits well in the existing enterprise model. Basically, we were building one offs that were difficult to support but were exactly what the customers were asking us for at the time.

What we learned very quickly from customers was that they wanted to work with us more once we expressed our strong opinions on what they should be building and running on their premise rather than letting them tell us what to build. This was the opposite of our initial assumptions that people wanted our expertise to build them a custom cloud – the reality is that they just wanted the Rackspace Cloud in their own data center.

In addition to being prescriptive about what a “good” private cloud looks like from a software and infrastructure perspective, we also believe strongly in the operations model of cloud. Last month I blogged about how your cloud needs a pit crew and how software and hardware makes up 25 percent of your cloud and good operations makes up the other 75 percent.

When customers say they want to “run like Rackspace,” what they really mean is that they want the experience of just being able to use their cloud. They want our bits, for sure – and those are free and open source with our OpenStack-powered Rackspace Private Cloud Software. But in addition to the software, customers want our expertise in configuring, operating, upgrading and the overall care and feeding of their private cloud. This is what we’re really good at. We do it for our public cloud, and we’re doing it now for a growing number of enterprises.

There were some Twitter and blog discussion last week about how the only thing we would operate is our own OpenStack-powered Rackspace Private Cloud Software. The biggest source of discussion that I saw was a question that went like this: “is Rackspace Private Cloud a fork of OpenStack and that’s why they won’t/can’t provide operations for other versions?”

Short answer: No, we deliver community OpenStack with a proven Rackspace Cloud configuration, tools and deployment model – not a fork. We work hard to be completely open and transparent in the community about what we’re doing.  Everything we do for Rackspace Private Cloud is on github, and Rackspace Private Cloud leverages the fully open source community packages for the OpenStack components (Nova, Swift, Glance, Horizon, Keystone, Cinder, Quantum, etc.) directly from the Ubuntu package repository. These are the same exact packages that the OpenStack developers themselves use for their development and test environments. Rackspace Private Cloud Software then deploys and configures those packages “the Rackspace way” using more open source tools and cookbooks that we’ve built (all available at http://github.com/rcbops/).

The reason we aren’t offering private cloud operations for other private cloud software is simply market demand – customers are asking to run like Rackspace, which is hard to do on someone else’s interpretation of what a cloud should look like (see: Crazy Cloud, above).

Don’t build Crazy Cloud!  Our Rackspace Private Cloud Software is free and open source, we have published reference architectures for private cloud that are open and freely available on our website and you don’t have to pay us for anything or worry you are getting a fork of OpenStack when you choose to do things the Rackspace way.

Remember the cloud equation: software and hardware are 25 percent, operations is 75 percent. If you truly want to run like Rackspace, there’s only one way to do that – let us use our experience in operating large OpenStack-powered open clouds to operate your private cloud so you can focus on the real value of cloud: using it.

Questions or comments about this blog post? Let me know below or on Twitter: @scottsanchez

There’s no doubt that OpenStack software and high quality compute, storage and network hardware are critical to having a winning cloud – but it’s a small part of the total “cloud” equation. Here’s what I’m calling “Sanchez’s Law of Cloud Mechanics:”

Winning with Cloud = 25 percent Great Software and Hardware + 75 percent Great Operations

I’ve been speaking to customers and partners about this equation for a while, and recently (maybe because of the F1 madness in my home city of Austin) I had an analogous epiphany: racing. Here’s the same math as it relates to auto racing:

Winning the Race = 25 percent Great Car and Driver + 75 percent Great Pit Crew

If I gave you the best car in the world and the best driver in the world, and six warm bodies from the stands that read the manual about the car before the race, you would never win. The pit crew is working both on and off the track before, during and after the race. The folks in the stand see very little of the work that goes on, but it’s absolutely critical to winning the race. A majority of what it takes to win a race is done off of the track. Just take a look at this short list of what a pit crew does:

• Pre-race strategy, planning and preparation to configure/tune/equip the car
• Engineering like aerodynamic testing, fuel and tire modeling
• Logistics, ordering, planning and supply chain management
• Real-time monitoring and feedback to the driver of key vehicle, track and environmental metrics
• Regular maintenance during scheduled pit stops – tires, fuel, performance adjustments – without unscheduled down time
• Critical repair after vehicle damage during the race – every second matters to get back in the race
• … and much more

The racing math is simple – without an amazing pit crew watching the vehicle and the driver, and doing the hard work before, during and after the race, you’d have zero wins. The pit crew that wins an F1 race doesn’t come from your corner oil change shop and work weekends on the F1 track as a part time job. These men and women spend decades working their way up. Maybe they start as the person that carries and changes the tires. Then they’re the person running big data-driven analytics on performance of a dozen different tire choices on 13-month asphalt in 57-degree weather with 31 percent relative humidity. This experience culminates and that person becomes the guy or gal making a critical decision that will affect the outcome of the race. This crew is the SEAL team of operations. Their dedication to behind-the-scenes excellence is why the driver can get in the car and perform his or her art and win the race.

This same mentality applies to running a cloud. The best software and hardware in the world won’t give you a winning cloud. You need an amazing group of cloud operators. It’s relatively easy to find cloud software and run it on some hardware. Your current admins are very likely not capable of running your cloud. Cloud is different. It’s never “off.” Scheduled downtime? Hah! Very few people have the tools, mindset or experience to effectively and efficiently operate a cloud, and you probably can’t afford to allow them to get that experience and expertise on your dime (like when your business critical app is down).

Rackspace is in the business of delivering Fanatical Support. We’re committed to delivering an amazing cloud outcome for our customers – in our data centers or yours. The (now not so) “secret” recipe is 25 percent Software and Hardware and 75 percent Operations. We founded OpenStack and helped it to become amazing software with a thriving community. But, we also learned the hard way what goes into amazing operations.

How did Rackspace become great at operating clouds? By doing it, at huge scale, across numerous data centers and continents for the better part of a decade. Fanatical Support is a culmination of experience and dedication that defines Rackspace and the 5,000 Rackers around the world who come into work each day to volunteer their best.

You can hire the folks from the local oil change shop to be your pit crew and hope for the best, or you can hire proven, experienced experts who live and breathe this stuff. Operations is our bread and butter, and doing it well is part of our DNA.

How badly do you want to win the race?

Questions or comments about this blog post? Let me know on Twitter @scottsanchez.

The market— vendors, analysts, etc— has conditioned buyers to think about the cloud as three or more separate projects and buying decisions. There are even more if they separate them by cloud vendor or technology choices.

Let me be clear:

Don’t think about your enterprise cloud strategy as different pools of resources from different providers or technologies. Build and deliver ONE cloud to your users.

Think about the cloud as the attributes it provides that you and your users care about. (Go back and read blog posts from the last decade if you aren’t sure what they are.) Then, plan how you will make the pools of resources with different characteristics available to your users in various locations and as transparently as possible. Do you need security? The cloud offers that, but the resources might be on-premise. Do you need to be in a specific geography? The cloud can do that, too, but it might be with a specific public provider. Do you need specific performance attributes? The cloud makes that possible, but it might be spread across on-prem and public resources to meet your needs. Your users need to see the cloud as one thing, and you need the governance to make that possible.

So when an enterprise user asks, “Where should I put this app?” your answer should be: “the cloud” – not “cloud No. 5.” You should have the tools and processes wrapped around your implementations to make sure that answer actually works.

Agree or disagree? Follow me on Twitter @scottsanchez and tell me!

First, Alamo is a compilation of everything you need to go from bare metal to a production OpenStack-powered private cloud in 30 minutes or less. Boot from the free ISO, enter some IP information and set a password, go get a coffee and you’ll have a running cloud before you know it. Rackspace made all of the decisions regarding the operating system, hypervisor, version of OpenStack, automation and deployment tools, and all of the configurations for those components. The software is available for free and Rackspace offers optional support for companies wishing to have our experts on-call 24×7.

Here’s the breakdown on what is contained in Alamo and what their specific licenses are. In addition to the included components and their individual licenses, Rackspace wrapped a copyright around the entire compilation the same way folks like Ubuntu and Red Hat do with Linux and the packages of open source software they compile. There is a simple end user license agreement from Rackspace that grants the user free access to the Rackspace-branded compilation, but it’s important to remember that all of the components are already free and open source and we do not place any additional restrictions on them.

So are the components open? Yes! Think about it like this: just like a musician that uses a guitar chord from an electric guitar and a piano stab from a grand piano to create a song that they copyright and distribute, Rackspace has done the same with Alamo. A musician does not restrict other people from using or sampling a piano note or a guitar chord because they used it in their song, just like Rackspace doesn’t restrict how people use the open source components of Alamo. The underlying licenses (GPL, Apache, etc.) define how and what people can do with those components. Although we’ve configured OpenStack to be deployed and usable, we haven’t forked the code. In fact, we’re using 100 percent “community” OpenStack code.

Bottom line: Rackspace Private Cloud Software uses open source components with large communities, and the end user license agreement from Rackspace for the compilation grants you the rights to use the compilation Rackspace has built and distributes. We have an end user license agreement to protect the Rackspace brand, and to protect the branded (but open source) software contained in Alamo that we redistribute.

Questions or comments? Let’s talk about it on Twitter: @scottsanchez