But that has nothing to do with the cloud, because the cloud is all about automation. Mainframes are about manual provisioning and management. IBM does have experience with automated provisioning of HPC clusters, but that's also fairly different from the cloud.
Is that true? I know nothing about mainframes, so it's an honest question
IBM often writes "performance test" periods into their contracts -- you get x weeks per year to trial the higher levels of performance. This is a sop that realises that mainframes are busiest at financial reporting times, especially at the end of financial year.
Which is a pity, because I look after a bunch of Oracle on zLinux and would like to share...
There are definitely places where IBM plays well, but commodity clouds aren't one of them, at least to my mind. If you need a high-end computing cluster that's bigger than the max Amazon instances, that would definitely be a place where IBM could play, it's just not a place I see SoftLayer playing.
You're right though; There is more than one way to skin a cat or a cloud.
The thing is that they don't want to cannabalise their own business. If they made zSystem mainframe linux instances accessible with a credit card (~2-5 seconds to boot, RAM-speed local networking, 5-9s hardware availability) they'd risk severely cutting into some of their most lucrative lines of business.
So they won't do that. Not right now, anyhow.
IBM currently quotes ability to present "thousands" of virtual Linux images on a single current zSeries mainframe. These servers have virtualised storage IO, virtualised network interfaces etc. and management interfaces older than most of the developers on the hipper cloud platforms.
So to claim IBM "can't build a cloud of their own" is quite ridiculous - they can deliver single mainframes with more capabilities than most smaller cloud providers entire infrastructure for many types of workloads, and have decades of experience aggregating workloads on collections of servers like this (and single system image clustering across up to 32 zSeries mainframes - some simple math should indicate you're then in the tens of thousands of VM images managed as one single system - this with software originally introduced in 1990....), as well as decades of experience building super-computing clusters with accompanying orchestration and management. Heck, they have a separate division doing integrated service management software.
Claiming they can't build a consumer level / low end cloud solution would be more believable, and would explain the SoftLayer acquisition. But there's far more to this market than going for startups who order a handful of vm's.
> When all you have is hosting, you can't lose money on hosting.
That might have become an issue for SoftLayer in the long run if they remained independent.
But IBM is a services company. And one that is an expert in extracting money from customers by building very deep and far reaching client relationships when you can afford them. They could lose money on hosting for the next century if they like, and still see it as a cheap foot in the door for their sales team if it helps them reach new customer segments.
> Google and Amazon are playing the ultra-thin margin long game and the result is that they hope to destroy the competition with an aggressive pricing floor.
Amazon at least is priced ludicrously high compared to providers like Softlayer that can offer hybrid "dedicated + cloud" solutions. If Amazon can compete on price in that market, they've so far not demonstrated any will to do so. Amazon is priced well for the batch "bring up ten servers for two hours" market, though.