
Gartner
Past experiences with Gartner forcecasts says that the eventual outcome is usually the exact opposite of what they predict.
Fresh from the news that Amazon Web Services intends to replace its hypervisor, the Xen Project will tell the world it has a fine future in embedded applications. Publicly-posted minutes of the Project's November Advisory Board call mentioned The Register's report of Amazon Web Services documents that revealed the cloud …
When I see KVM I think of all those benchmarks where it was outperformed by Hyper-V!
There's something to be said for licensing costs, you know. Take it from somebody migrating from VMware to Hyper-V for that very reason.
If I could get Veeam on KVM then there's every chance I'd go that way...
"There's something to be said for licensing costs, you know. Take it from somebody migrating from VMware to Hyper-V for that very reason."
You mean as in that Hyper-V Server is completely free with all features enabled versus the full enterprise version of Redhat at over $9K per server per year?
Tried to post the link of the source quoted below, but the comment did not get published. Not sure whether this is an editorial decision or maybe the moderator is just on vacation for Thanks Giving.
In any case, the source can be found if you google news for "xen". And it could have been found by the author of this article also.
<Quote>
AWS isn’t abandoning Xen, its hypervisor of choice since the outset of the platform. Instead, it will adopt a multi-hypervisor strategy with both Xen and KVM depending on a given workload’s specific requirements. We asked AWS if the introduction of KVM had to do with any issues with Xen; an AWS spokesperson responded with a statement that the P3 instances on sale since October use Xen, and the company will continue to heavily invest in Xen.
“For future platforms, we will use the best virtualization technology for each specific platform and plan to continue to launch platforms that are built on both Xen and our new hypervisor going forward,” the spokesperson said.
<End Quote>
See http://searchaws.techtarget.com/blog/AWS-Cloud-Cover/AWS-using-KVM-and-Xen-but-users-may-not-feel-any-impact
<Quote>
AWS isn’t abandoning Xen, its hypervisor of choice since the outset of the platform. Instead, it will adopt a multi-hypervisor strategy with both Xen and KVM depending on a given workload’s specific requirements. We asked AWS if the introduction of KVM had to do with any issues with Xen; an AWS spokesperson responded with a statement that the P3 instances on sale since October use Xen, and the company will continue to heavily invest in Xen.
“For future platforms, we will use the best virtualization technology for each specific platform and plan to continue to launch platforms that are built on both Xen and our new hypervisor going forward,” the spokesperson said.
<End Quote>
I think the point is that you are contradicted by an official statement from the company in question in response to your "scoop".
Which is important because you yourselves are using the lack of a contrary statement as evidence that the original story was correct; "plenty of time for AWS to request a correction if our reports are not correct." An alternative interpretation is that AWS, like Apple, consider El Reg to be an irrelevance (and any number of other explanations, like the press office being on holiday, busy preparing for reInvent etc.)
Time will tell who's right.
We reported exactly what AWS published and none of that mentioned multi-hypervisor strategies. When we saw AWS' posts we immediately asked AWS for comment. As it happens, AWS immediately reached out to ask us for a correction of an unrelated error in the piece (I mixed up cores and threads). So there goes the "AWS is too busy and you don't matter to it" argument.
They know where to find us and are not shy about doing so.
But AWS never asked for any change to either piece regarding the hypervisor content. I was not aware of the TechTarget piece until today and will refer to it now.
We don't aim to be relevant to vendors. We aim to be relevant to readers. And I'm pretty sure they appreciated reading about a significant piece of news here, first, TWO WEEKS before TechTarget touched it.
The TechTarget story doesn't contradict the Register article. AWS are continuing to invest in Xen as the PR announcement says. No one believes that suddenly AWS is going to migrate thousands of VMs to KVM.
They'll probably do things like run Xen nested on KVM for backwards compatibility. Linux for hardware support and Xen for APIs.
As the Citrix guy said: none of their customers care about hypervisor layers any more. They just use whatever is in Azure or AWS.
Nice of them to tell that to Techtarget - as our story states, when we read of the new KVM hypervizor, we immediately asked AWS for an interview about their strategy. And they never replied. Further, their statements to TechTarget contradict the words in their own FAQ, which said all new instance types would use KVM. See our second story for the screen grabs to that effect
I agree with AC above: there is no contradiction.
I'm a Linode customer, and it was a royal pain rebooting servers every couple of months because yet another security hole was found in Xen. Linode abandoned Xen and the world is a better place: https://www.cio.com/article/2937714/cloud-computing/why-linode-moved-to-kvm.html .
My take on this story... AWS have come to the same conclusion: Xen is a dud. As general policy, they will create new images based on KVM. For now, they will not disrupt the stack of existing images, and minor variations could still use Xen. I suspect the P3 is an example of the latter.
"We asked AWS if the introduction of KVM had to do with any issues with Xen; an AWS spokesperson responded with a statement that the P3 instances on sale since October use Xen, and the company will continue to heavily invest in Xen."
Note the spokesperson did not directly answer the question. "Heavily" means whatever I want it to mean.
We must speculate because Amazon will not give us a definitive answer. They have no need to broadcast in detail what they're doing. Most customers don't care. Over-detailed descriptions give competitors a shortcut - instead of doing their own research and testing, they can just copy AWS. A bit like the Scottish fast food chain doing intensive research on the ideal site for a store, and the competition simply opening up next door.
AWS will be supporting and indeed fixing Xen for the foreseeable future, so it would be stupid to disengage from the Xen project. Jeff Bezos and the gang are not stupid. Weaning themselves off Xen will take years, and will happen by natural attrition anyway as they introduce new, more capable and more cost-effective images. No need to make a song and dance about it.
A year from now, Xen might be improved and AWS might revisit this decision. I doubt they will change their minds. No, the really significant thing is *letting KVM into the stack at all*. That's the real story, and Simon Sharwood is spot-on to notice. A non-answer from a PR person in Amazon does not contradict his story.
And can I say one of the reasons I read and like The Register is their bulldust detectors are pretty good. I think being based in the UK and not the US helps. There is a level of scepticism, even cynicism, when other IT publications (*cough* TechTarget *cough*) are too ready to copy-paste PR guff.
See this story: https://www.forbes.com/sites/janakirammsv/2019/03/10/how-an-acquisition-made-by-amazon-in-2016-became-companys-secret-sauce/#7e61c6632f67
AWS acquired Israeli company Annapurna which makes custom ASICs that dramatically speed up I/O for virtual machines. KVM was important to make it all work. So Simon Sharwood was on the money to report on this in November 2017.