HPE and Rackspace announce pay-as-you-go OpenStack private cloud

On 13 November 2017, Hewlett Packard Enterprise (HPE) and Rackspace announced a partnership to offer OpenStack private cloud on pay-per-use infrastructure. The companies called it the industry’s first pay-as-you-go OpenStack private cloud, delivered as a managed service.

The idea was to give enterprises public-cloud traits, namely utility-style pricing, elastic infrastructure and simpler IT, inside a private environment. That environment could sit in the customer’s own datacentre, in a colocation facility or in a datacentre run by Rackspace.

Who did what

HPE supplied the hardware. According to SiliconANGLE, that included its latest Gen10 servers, plus hardware support. Rackspace ran the private clouds for customers. Light Reading reported that, depending on the contract, either company would be responsible for fixing or upgrading the underlying hardware.

Metering and billing relied on HPE Flexible Capacity. Under that model, customers paid for what they used rather than for fixed capacity, and could absorb bursts and traffic spikes without buying extra permanent capacity. Constellation Research noted that HPE had acquired Cloud Cruiser that year. HPE had already been the largest customer of Cloud Cruiser’s usage-monitoring technology, which it had rebadged as Flexible Capacity.

Why both companies moved

Rackspace had been shifting away from its hosting roots towards specialised managed services, particularly for OpenStack, which it co-created with NASA in 2010, Constellation Research said. HPE Flexible Capacity gave it a consumption-priced hardware base to pair with that expertise. Scott Crenshaw, Rackspace’s executive vice president of private cloud, said the model removed barriers to private cloud adoption. He also said Rackspace planned to extend it to its other private cloud offerings.

HPE also gained a managed-service channel for its installed base. IDC’s Michelle Bailey said Rackspace was well placed to serve HPE’s large base of infrastructure customers and to reduce migration risk. She described pay-per-use as a step towards better flexibility and cost transparency for private clouds.

The claims

The companies said customers typically saved 40 percent or more against the leading public cloud. Silicon UK reported that the estimate rested on Rackspace’s internal pricing analysis. The comparison was the vendors’ own, not an independent benchmark.

Other claims included:

  • A single-tenant design meant to remove “noisy neighbour” performance problems found in multi-tenant environments, and to make security, compliance and data sovereignty requirements easier to meet.
  • Instant scaling of private cloud capacity, in a public-cloud-like manner.
  • A 99.99 percent API uptime guarantee from Rackspace, offered as part of the managed service.

The announcement also cited IDC’s FutureScape: Worldwide Datacenter 2017 Predictions. That forecast said pay-as-you-go models would account for 50 percent of on-premises and off-premises physical IT and datacentre asset spending by 2018.

Availability and what followed

The service was due to be generally available in all regions from 28 November 2017. Rackspace said versions of its private cloud powered by VMware and by Microsoft Azure Stack were expected in 2018.

Forbes contributor Patrick Moorhead described the offer as a mix of private-cloud control and public-cloud economics. He said the choice of where to run workloads would be a multivariate assessment of cost, control, simplicity, security, performance and flexibility.

In April 2025, Rackspace Technology announced a separate service, Rackspace OpenStack Flex. It described this as an on-demand, shared cloud service for infrastructure-as-a-service, with 24×7 expert support.