Why? it's IBM, one would think they'd put it all on one/more of their mainframes.. maybe consolidating onto mainframes isn't as cost effective as their marketing suggests..
Big Blue wants developers to create and test their code on its IBM Cloud, and it expects them to code for Red Hat's commercial implementation of the open source KVM hypervisor for x64 servers. Last June, IBM booted out the first of its CloudBurst cloudy infrastructure packages of servers, storage and the systems software to …
Mainframes are usually obscenely expensive to buy and support compared to x86 tin, not to mention the vendor lock-in issue as well. However, I'm surprised IBM don't push Mainframes more - after all, these points are surely a blessing for IBM, once they get their foot in the door (big discounts etc at first).
IBM wants a cut on any infrastructure build-out in the world, no matter what ideology or purpose stands behind it. If the current market trend is to build clouds, IBM will happily do it, giving to the partners the parts it is not convenient with (competence-wise or cost-wise) at the moment. In the long term, if the business turns out to be sustainable, IBM will acquire the crucial partners. Remember also IBM has no choice, in a way - it needs to keep acquiring current competence to go head to head with other market players to stay relevant and get the upper hand in hard sales talks with customers.
As for mainframes, this all comes down to maturity of the needs. If the cloudy apps become crucial for the businesses thet deploy them IBM can easily migrate them onto "Linux on mainframe" type of infrastructure. Notice, for IBM the very process of migration is also a business. So, low and medium importance apps can go to the general cloud and critical apps will eventually go to mainframe cloud.
Biting the hand that feeds IT © 1998–2020