IBM's had the Linux Technology Center (LTC) for a long time and has been contributing to the community. All the platforms Z, POWER etc... support Linux as a first class citizen and plenty of other ecosystems are also supported (i.e GCC, OpenJDK, etc...)
Maybe its time to re-evaluate the old biases? The old incumbents like Microsoft have warmed up to OSS, not sure why Big-Blue is getting this much flak.
Microsoft has been doing a lot to rebuild itself as a cloud provider.
IBM has been doing a lot to go out of business.
Microsoft's first CEO is still chairman and helping lead the company even from the sidelines.
IBM is a floating raft of failed leadership.
Microsoft isn't trusted fully by the community but they are making inroads under their new CEO.
IBM has been firing its most senior people in an effort to slow its cash burn and to hire younger folks. IBM also claims it's also to bring on folks with more relevant skills to emerging technologies. I think there is a lawsuit about this.
Anyway, IBM has done nothing in recent years to show they are a true contender.
If the Redhat team is able to pull a Next here and assume leadership roles inside of IBM this could be stellar. Big Blue's formidable sales team and reputation with a great product line overseen by passionate people would be powerful.
If the IBM existing leadership team emerges as the winners here it will likely continue to fade into irrelevance.
"Microsoft's first CEO is still chairman and helping lead the company even from the sidelines." this is not true? maybe you meant he's a technical adviser
Thing is, IBM does not have a great reputation for how it manages its staff as it tends to favour moving positions to low cost locations and redundancies that provide the minimum possible payouts
I've followed POWER for a while now, and it seems to be mostly marketing hype. The POWER8 was not much better than the Xeons out, and had almost no availability. POWER9 excels at a lot of IO-bound tasks, but again, it's nearly impossible to find. It doesn't help that there are very few systems you can buy out there. On paper, the POWER9 looks awesome. Then you look closer at the actual architecture and realize that a lot of pieces of the design are weird and/or seem somewhat useless. It also doesn't help that they're extremely expensive, especially since there are so few options.
So while I'd like IBM to compete with Intel, they need to pony up more money if they really want to push the industry. Don't make it so hard to buy one, and publish benchmarks/comparisons with Intel.
Based on your subcomment about evaluation, I'm very interested to hear your opinions about Talos' offerings - up to now I've only heard anecdotes from individuals drinking the security+ownership kool-aid.
What IBM offers is probably a lot more coherent and takes better advantage of what the POWER architecture has to offer, such as larger quantities of RAM, the per-node interconnect fabric, faster I/O (not just PCIe), etc (admittedly totally naive here). Plus of course there's z/OS, which I know enough about to respect (and want to play with someday :) ).
Talos basically offers only Linux and a mildly DIY standing-up experience (https://tenfourfox.blogspot.com/2018/05/a-semi-review-of-rap...), although this is likely to be reasonably painless for non-desktop configurations (and perhaps volume orders can come preconfigured).
As a bit of a pet idea I kind of want to colocate one of the 2U or 4U systems for generic web serving and similar duties, but I fear that running a blog/discussion system on such a machine may result in a constant effort (on my part) at keeping discussion focused on the "it's a different architecture, what comp-sci interesting things can we do with it" aspect instead of getting distracted by shallow OCD-meta-security bikeshedding.
(It's kind of sad that the collective consensus about new/different architectures has to always be about security nowadays, and not about unbiased exploration, which is what we're best at)
My impression is that POWER is more suitable for certain types of scientific computation that require high degrees of precision rather than for general purpose enterprise use. POWER9 is the only CPU architecture out there that has support for quad-precision floats. Alas, no CPU exists with support for octuple floats, since only astrophysicists need that level of precision.
Right, and that's what we were evaluating it for. It's also the only one with pcie4, 8 memory channels, and CAPI. Nonetheless, it needs software and libraries to be updated to use some of these, and in many cases, they're not.
sir, you make a good point. the world always changes. microsoft has changed, perhaps IBM has also. one who believes in sin and redemption perhaps should give IBM a chance.
To add some substance to my comment. I’ve interacted with IBM in different scenarios. I’ve had to manage a team of IBM consultants at work, very frustrating experience. I still shed tears every time I think the hourly rate paid vs the value they provided. When Watson was being hipped the company I was working for was approached to use the tech and come up with some PoCs, the dissonance between the biz dev pitch and the actual “solution” was abismal. Currently I’m at a different company and we have IBM as one of our customers, the image that comes to mind is a headless chicken running around.
That is to say, not a very impressive impression. I’ve to say that I’ve seen some interesting stuff as well. But I’m still very sad about the buyout
IBM's had the Linux Technology Center (LTC) for a long time and has been contributing to the community. All the platforms Z, POWER etc... support Linux as a first class citizen and plenty of other ecosystems are also supported (i.e GCC, OpenJDK, etc...)
Maybe its time to re-evaluate the old biases? The old incumbents like Microsoft have warmed up to OSS, not sure why Big-Blue is getting this much flak.