Having earlier this month rolled out the Bluemix Local platform as a service for firms to run behind their own firewall, IBM today has taken the wraps off an OpenStack on-premise private cloud service - a fruit of its June acquisition of Blue Box.
The new on-premise environment, IBM Blue Box Local, will absorb features from the existing IBM Cloud OpenStack Services offering which it replaces, when it becomes available from November 30.
With IBM Blue Box Local, after a site survey to determine the deployment capacity required, the customer buys the hardware as prescribed in a compatibility list. IBM then builds the rack and delivers it to the customer site. Once the power and network connections are made, the entire environment is then managed remotely by Blue Box and IBM.
"If you look at most private-cloud installations, they're measured in months or years to get up and running," IBM Blue Box CTO Jesse Proudman said.
"There are sites where typically the customer goes and buys a software distribution from a vendor, and they have to figure out how to train their people in that software distribution. They have to figure out how to operate that environment - what do they do when they get a monitoring alert?
"It becomes challenging to derive immediate value from that cloud.
"We're trying to take all that out of the mix and put a small portion of the IBM cloud footprint in that customer site," he said.
Having been in business since 2003, Blue Box created its OpenStack-based hosted private cloud some two years ago. Before the IBM acquisition, that product ran on dedicated hardware in four Blue Box datacenters around the world.
Following the IBM purchase, the first jointly-released product was the porting of that dedicated offering onto IBM's Softlayer cloud in early September. IBM Blue Box Local is essentially that same technology but placed within a customer's datacenter.
"There are customers that are OK with public clouds and they're typically doing net new workloads and putting those brand-new workloads in public clouds," Proudman said.
"Then there are customers that have significant existing investment in systems or database technologies but they want to begin to develop their next generation of applications with proximity or adjacent to that existing workload."
Typical examples of those types of companies are to be found in the healthcare or banking industries. They want their new applications to talk to their systems of record, which they are reluctant to move to a public cloud.
"They don't want to write those applications using the same development legacy practices they [used when they] wrote those applications maybe five or 10 or 20 years ago. They want to write them in a cloud-native format but they need to be able to interconnect that with an existing implementation," Proudman said.
"That customer doesn't want to have to go hire an OpenStack team, install OpenStack, and figure out how to operate it. They just want a cloud platform that is within proximity to what they're trying to interface with.
"This product brings that to them. Imagine taking all the pain out of OpenStack and just delivering the experience of OpenStack to that customer."
Proudman said OpenStack is a phenomenal set of technologies but at the same time it is incredibly complicated and is only becoming more complicated with time.
"It's so difficult to find application engineers, let alone application engineers who are familiar with OpenStack and all its technologies.
We can centralise that expertise into one organisation," he said.
"As one organisation managing thousands of these clouds, we have this attribute where from each set of learning we get when one cloud fails we figure out what caused that to fail and then we take that back into the product. Then we're able distribute that back into all the environments in real time because we're managing them.
"So you get this network effect where every additional cloud sold or every capacity block that's added, the whole aggregate platform, every customer's experience gets better and more reliable."
The OpenStack open-source project was started in 2010 by Rackspace and NASA to create components for building public and private clouds on standard hardware.
It is now backed by more than 540 supporting firms and vendors, including Cisco, Dell, HP, IBM, Intel, Oracle, Red Hat, and VMware, with a large developer community working on a range of loosely-coupled projects.
Read more about IBM and OpenStack