Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Microsoft and Red Hat partner

blogs.microsoft.com

Post reply on HN
Print view View original

101–110 of 177 posts

Re: Microsoft and Red Hat partner

#101

Post by scholia »

toyg wrote: post #64
scholia wrote: post #48

Earlier quoted context omitted.

Red Hat could have been one of the first Linuxes on Azure, if it had wanted. But back in 2013, it said: "Red Hat CEO: We don't need Microsoft to succeed" http://www.infoworld.com/article/2614357/linux/red-hat-ceo--... I don't see why you're quibbling about "late". It's a fact that it's more than three years later than a bunch of other versions. But hey, you can have your own spin.

Note how RedHat didn't say that: if you read the actual article (despite being from InfoWorld): > InfoWorld: Microsoft has a close business relationship with Suse Linux. That seems to be Microsoft's Linux of choice, and the company doesn't seem interested in having the same kind of partnership with Red Hat. Is that a problem for Red Hat? > Jim Whitehurst: We'd be happy to work on interoperability with Microsoft or an…

I know Microsoft pretty well, and it would have wanted to support as many versions as possible, within the available time-frame. It obviously didn't just support SuSE, so your implied either/or is just yet more of your spin.

I don't know Red Hat that well (it's a few years since we've talked), but it has a strong focus on its own cloud business (1). It might have seen Azure as a rival to Red Hat cloud services, but that's just my speculation.

(1) https://www.redhat.com/en/technologies/cloud-computing

BELOW

> I have no horse in this, no investment, nothing; I just don't like unsupported bias. Can you say the same?

If you really don't like unsupported bias, perhaps you shouldn't post comments that reveal so much of it ;-)

> Yes, it is "just your speculation".

You could also cut out the cheap tricks. I referred specifically to the comment on Red Hat's motives, not to anything else.

Re: Microsoft and Red Hat partner

#103

Post by Alupis »

thebouv wrote: post #88
baldfat wrote: post #47

Earlier quoted context omitted.

Azure has been mostly Linux Applications for a while now. Whenever Enterprise wants a contract they almost exclusively think Red Hat. So this makes sense for both of them to team up.I also don't see any fallout in the community because of this. Since Ubuntu ruling the cloud instances this also makes sense for Red Hat.

Ubuntu rules cloud instances elsewhere? I almost always see the choices as CentOS or RedHat. But that might be my myopic view cause I am always just looking for those right off the bat.

Ubuntu was made the default instance on EC2 (even though EC2 is built on RHEL/CentOS/XenServer).

That's were Canonical began to gain traction regarding installs, yet they still can't find a way to get users into support contracts en masse like Red Hat has been able to.

Re: Microsoft and Red Hat partner

#104

Post by aexaey »

X-Istence wrote: post #98
aexaey wrote: post #51

Earlier quoted context omitted.

I'm getting approximately the same CPU & storage performance out of Azure's A0 machine w/Debian8, as from AWS t2.micro instance with spinning rust storage and Debian8 (both are cheapest offer available, and price is quite similar between them). Network on Azure though is a totally wrecked thing: - A0 instance is capped to a miserable 5 megabit/s uplink (downlink is fine); - If you want TCP/UDP ports open, you need to…

I don't see why private IP's in a trace route is necessarily wrong... Many large providers are numbering their hops with private IP space and routing public space across it because they don't want to waste /30's of public space for each point to point. Yes, /31 is an option, but there were a lot of edge cases or network gear that didn't support it.

Yes, you are absolutely right that IPv4 addresses are scarce.

Which bring a related question - what's up with none of AWS/Azure/GCE supporting IPv6?

Re: Microsoft and Red Hat partner

#105

Post by astine »

ceejayoz wrote: post #15
cdnsteve wrote: post #6

A cloud provider that offers Linux and Windows, this is cutting edge.

Sure, it's not cutting edge, but it's one of a bunch of signs recently that Microsoft has turned over a new leaf in the last few years. Visual Studio's page now mentions Git, a GPLed piece of software built for Linux kernel development. They've come a long way from "GPL is cancer!"

Git is actually integrated in the latest VS and a lot of recent documentation assume that you are using it by default. Also, the latest release of ASP.NET MVC uses Bootstrap as its front end. That's not GPL, but it seems to be part of the larger trend of Microsoft integrating popular FOSS tools into their workflow where they can.

Re: Microsoft and Red Hat partner

#106

Post by evan_miller »

Someone1234 wrote: post #63
diakritikal wrote: post #17

In the last 5 years we've experimented with Linux on top of hyper-v a few times. The basic I/O, compile & runtime performance was significantly inferior to xen & kvm (didn't bench against vmware, phasing it out due to cost), it wasn't worth any effort to even deploy apps for testing. Therefore I don't see Linux on hyper-v being a compelling option for the cost conscientious technical officer or lead engineer.

With all due respect: I don't buy it. That isn't how the technology works. Compilation in particular would never even hit the hypervisor layer regardless of which one you were utilising. It is CPU/memory bound. The difference between hypervisors for IO is tiny. They're all slower than native, but vary very little between one another. That's why the manufacturers have all almost given up trying to use performance for…

You're missing something very significant: Compilation jobs involve a lot of forking. Forking, memory accesses, and TLB differences across various VM architectures can have an enormous impact on performance.

For example, we found two orders of magnitude performance difference between PV and HVM on ec2 on fork-heavy CI jobs. A difference between 3 and 30 minutes. This difference is trivially reproducible by measuring the time taken to complete a fork() syscall. As the memory size of the parent process grows, the difference becomes more pronounced.

Here's actual data I collected from c3.large instances, using a contrived Ruby one-liner to benchmark fork():

PV: [ec2-user@PV ~]$ ruby -e'$ref=[]; (24..29).each {|n| $ref n); start = Time.now.to_f; fork.nil? and exit!; puts Time.now.to_f - start; sleep 1}'

0.00550532341003418

0.011992692947387695

0.02542281150817871

0.05213499069213867

0.10562920570373535

0.21283364295959473

HVM: [ec2-user@HVM ~]$ ruby -e'$ref=[]; (24..29).each {|n| $ref n); start = Time.now.to_f; fork.nil? and exit!; puts Time.now.to_f - start; sleep 1}'

0.0005323886871337891

0.0007219314575195312

0.0012335777282714844

0.002183198928833008

0.004210233688354492

0.008102178573608398

As you can see, the difference is considerable. This is all on Amazon ec2, the only difference between these VMs is PV versus HVM.

(edit: HN formatting seems to be eating the exponent operator in the "2 to the n" expression. It's a double *)

Re: Microsoft and Red Hat partner

#107

Post by mobiplayer »

aexaey wrote: post #51
diakritikal wrote: post #17

In the last 5 years we've experimented with Linux on top of hyper-v a few times. The basic I/O, compile & runtime performance was significantly inferior to xen & kvm (didn't bench against vmware, phasing it out due to cost), it wasn't worth any effort to even deploy apps for testing. Therefore I don't see Linux on hyper-v being a compelling option for the cost conscientious technical officer or lead engineer.

I'm getting approximately the same CPU & storage performance out of Azure's A0 machine w/Debian8, as from AWS t2.micro instance with spinning rust storage and Debian8 (both are cheapest offer available, and price is quite similar between them). Network on Azure though is a totally wrecked thing: - A0 instance is capped to a miserable 5 megabit/s uplink (downlink is fine); - If you want TCP/UDP ports open, you need to…

Hi aexaey,

There are some things wrong on your comment, please let me comment on them:

I don't think A0 are capped at all, I don't know where did you get that:

user@A0VM:~$ speedtest-cli

Retrieving speedtest.net configuration...

Retrieving speedtest.net server list...

Testing from Microsoft Corporation (40.113.XX.XX)...

Selecting best server based on latency...

Hosted by KsFiberNet (Wichita, KS) [45.26 km]: 139.703 ms

Testing download speed........................................

Download: 44.48 Mbit/s

Testing upload speed..................................................

Upload: 22.54 Mbit/s

I just had to run one single test to get these results. Multi-threading will probably pump this up (let's say a multi-thread iPerf test between two VMs on two different regions, through their public IP addresses).

(Funnily enough this VM is in Dublin, hence the latency to Wichita).

You can get access to all your TCP/UDP ports by adding an instance-level IP: https://azure.microsoft.com/en-gb/documentation/articles/vir... (you can later protect it with NSGs --> https://azure.microsoft.com/en-us/documentation/articles/vir... ). What you've seen is when you create endpoints on the load balancer that points to the VMs behind it. Now in Azure Resource Manager, spinning up a VM from the new portal will actually get you an ILPIP and no load balancer by default.

About that statement about a ton of weirdly configured network gear: Well, for starters you're doing network virtualization here :) but seriously now, the platform doesn't forward ICMPs to/from Internet, so I'm curious about those private IP addresses on a traceroute (and of course the packet drops). If you're talking about hops between you and the VIP (the load balanced public IP in front of your VM), that's probably your provider, other providers and finally Microsoft's Network. As per the rest of the path, you've said that ICMP doesn't go through, so traceroute it's not going to work (tcptraceroute also needs ICMP to work :)).

Re: Microsoft and Red Hat partner

#108

Post by scholia »

toyg wrote: post #72
lloydde wrote: post #25

Earlier quoted context omitted.

Worked out well for Xamarin. Also, it seems to me that SUSE is still in good shape, but didn't continued to grow like RedHat.

> Worked out well for Xamarin Only for them. Novell bet big on pivoting on Linux and then crashed.

Eric Schmidt on Novell, recently:

"I went to Novell under the mistaken goal of being a CEO. I didn’t do the due diligence, and if I had, I wouldn’t have gone. Our basic goal was to get out with our professional reputations intact and not end up in jail. The books were cooked, and people were frauds. But it turns out you can overcome that, and the skills I developed helped at Google."

https://medium.com/cs183c-blitzscaling-class-collection/cs18...

Re: Microsoft and Red Hat partner

#109

Post by scholia »

Eric_WVGG wrote: post #50
yxhuvud wrote: post #3

Now that is a headline that would have raised a few eyebrows 15 years ago.

I’m definitely picturing Balmer quietly fuming, pacing a bit, staring at a Clippers logo and thinking “eye on the ball, Steve, eye on the ball”

Microsoft started offering multiple versions of Linux on Azure while Ballmer was in charge.

Microsoft also started writing dozens of apps for Android and iOS while Ballmer was in charge.

Re: Microsoft and Red Hat partner

#110

Post by scott_karana »

Alupis wrote: post #103
thebouv wrote: post #88

Earlier quoted context omitted.

Ubuntu rules cloud instances elsewhere? I almost always see the choices as CentOS or RedHat. But that might be my myopic view cause I am always just looking for those right off the bat.

Ubuntu was made the default instance on EC2 (even though EC2 is built on RHEL/CentOS/XenServer). That's were Canonical began to gain traction regarding installs, yet they still can't find a way to get users into support contracts en masse like Red Hat has been able to.

> Ubuntu was made the default instance on EC2

Citation needed. If Amazon Linux is still the default, you're wrong. It is based on CentOS, uses RPMs and yum.

Post reply on HN