Disclaimer: my company was acquired by Rackspace 2 years ago. I am a developer myself and below are my honest impressions (and things I have learend) of Rackspace cloud.<p>Back when they launched it after having had acquired Slicehost, Rackspace Cloud was the only alternative to AWS and had one significant differentiator back then: the local storage on the instances wasn't ephemeral - it was permanent. The benchmarks and pricing were competitive, so the growth was phenomenal and life was good as a result. And that's when Rackspace dropped the ball by losing some focus and getting obsessed with launching platform services one after another ignoring the compute. They even launched what I believe was the first PaaS ever and gave it an unfortunate name of "cloud sites" (and confusing plenty of developers at that point, who couldn't tell "cloud sites", significantly less powerful service, apart from "cloud servers")<p>As a result, for a few years the lineup of VMs they offered hasn't changed, while AWS was launching new types all the time.<p>So about two years ago Rackspace ended up with a great control panel (easiest cloud CP to use IMO), a comprehensive set of back-end services: block storage, object storage, MySQL-as-a-Service, email sending and receiving, MongoDB-as-a-Service, cloud deployments, CDN and many more... But the selection of virtual servers was small and the hardware was getting obsolete at that point.<p>Since then I've seen tremendous investments being made into the core: the compute. Rackspace launched "performance" line of VMs. They still feature persistent (backed by RAID-10 SSDs) local storage, latest Xeons and ridiculously performant network. In a way it's "advantages of backwardness" at work: the performance line of compute is all fresh latest&greatest gear: the servers, the networks, etc. The performance, particularly on disk I/O, is awesome. I hope they'll keep the ball rolling by keeping it up to date by launching new generations in time.<p>Moreover, Rackspace hired (or acquired) some brainy engineers out of Bay Area who've pushed for "at scale" compute offering which is now called OnMetal (<a href="http://rackspace.com/onmetal" rel="nofollow">http://rackspace.com/onmetal</a>) - because that's how internet giants internally do compute, container-based architectures without virtualization, running on highly specialized high-uptime servers (open compute). So... Rackspace cloud today is the only way to experience "facebook-style" infrastructure by renting it by the minute.<p>The culture of "true web scale" infrastructure is very strong internally, that's why Rackspace has been a huge supporter and believer of Docker, and some Rackers - having grown tired of the sad state of Linux distribution - founded CoreOS (<a href="http://coreos.com" rel="nofollow">http://coreos.com</a>), the fresh look at running massively scalable infrastructure from the operating system point of view. Today, Rackspace is the only cloud where you can run Docker containers on top of CoreOS natively on highly efficient (performance per dollar) instantly-provisioned OpenCompute gear without noisy cpu/disk/network neighbors and without virtualization tax.<p>What could be better?<p>Well... Rackspace is not the cheapest for undersized deployments [1]. They've never been known for doing things like selling the over-provisioned VMs running on the cheapest machines on consumer-grade SSDs in RAID-5 arrays. A typical Rackspace customer treats a server outage as a big deal, "engineering for failure" for many (most?) cases is actually more expensive than just outsourcing reliability into a higher quality infrastructure. They love to make everything redundant and tend to go with higher spec and high-MTBF parts. Another component of the higher pricing is the support organization. The culture of bending backwards to make customers happy is insane, here's an example: if a customer asks for help resolving problems with Sendgrid (a competing email sending service), Rackspace support people will gladly assist, without bothering them with showing off the superiority of Mailgun - my company Rackspace acquired. I was blown away by that, customer gets help with whatever they feel like getting it.<p>You can even go as far as asking Rackspace to just run certain components of your infrastructure for you, saving on Devops positions. That's their bread&butter.<p>While Rackspace offers alternatives to most AWS services, they're not 1-to-1 equivalents. CBS (alternative to EBS) is vastly superior in my opinion because (based on what I know about EBS from public sources) CBS is simpler, has been delivering more predictable and higher performance/uptime experience, Cloud Files (S3 alternative) is comparable, ObjectRocket (MongoDB-as-a-Service) is absolutely phenomenal - the only DBaaS I can recommend anywhere, but there're no Redshift or Glacier counterparts.<p>I guess what I'm saying is this: Rackspace Cloud has evolved into a grown up contender with a solid feature set matched only by AWS, with multiple DCs all over the world and some unique capabilities (like OnMetal flavor of cloud servers), which differentiates by over-investing into hardware infrastructure and support. So if those qualities matter, take a look. The program Jesse is announcing only makes it easier for you.<p>[1] Once you cross $10K/mo hosting spend, they're very competitive especially when you factor in the cost of running/managing the infrastructure.