Microsoft Azure Overtakes Amazon's Cloud In Performance Test 82
alphadogg writes "Microsoft Azure's cloud outperformed Amazon Web Services in a series of rigorous tests conducted by Nasuni, a storage vendor that annually benchmarks cloud service providers (CSPs). Nasuni uses public cloud resources in its enterprise storage offering, so each year the company conducts a series of rigorous tests on the top CSPs' clouds in an effort to see which companies offer the best performing, most reliable infrastructure. Last year, Amazon Web Services' cloud came out on top, but this year Microsoft Azure outperformed AWS in performance and reliability measures. AWS is still better at handling extra-large storage volumes, while Nasuni found that the two OpenStack powered clouds it tested — from HP and Rackspace — were lacking, particularly at larger scales."
Web scale! (Score:2, Funny)
So Microsoft Azure is now web scale? Does it use MongoDB?
Re: (Score:3)
I think you missed the reference [xtranormal.com]. If you haven't seen it before, it's hilarious.
Re: (Score:2, Funny)
Yeah, but fuck Microsoft because stuff.
Re: (Score:1)
Re: (Score:2)
What exactly is there about FOSS that you need to "get the scaling benefits of the cloud"?
(SQL Azure is one of the storage options for Azure, but it's not the only one, and it's specifically the one that sacrifices scalability for the sake of retaining relational features. If you want scalability, you should be using Table Storage, which is basically the usual "NoSQL" key-indexed table thingy with no complex queries or joins).
Re: (Score:1)
What exactly is there about FOSS that you need to "get the scaling benefits of the cloud"?
a) Ability to use 10X the amount of software installs and 10X the amount of virtual machines that you have ever used because today your special offer went viral on Slashdot.
b) Strong guarantees of future portability and availability on all different platforms by using an open source abstraction layer; either CentOS/Red Hat at the IaaS level or Python/Ruby/Java at the PaaS level where you know there is no single owner proprietary vendor that can stop the technology being ported to different cloud systems i
Re: (Score:1)
a.1) and not getting arrested because just your software control software went a bit beserk and actually used twice as many OS instances at one moment as you thought you did and declared. When things go all dynamic they can get a bit weird. You do not want to have to explain that to a BSA audit crew.
In unrelated news... (Score:5, Interesting)
Nasuni scores $20M to build out storage sales effort
Nasuni, which helps distributed companies manage their cloud storage securely, has $20 million in a new funding round — led by a mystery investor — to help it pay for new features and expand sales and marketing, said CEO Andres Rodriguez.
http://gigaom.com/2012/10/30/nasuni-scores-20m-to-build-out-storage-sales-effort/ [gigaom.com]
Re: (Score:2)
Yes but the thing is cloud environments are a shared medium.
If you just happen to get a node where someone is calculating Pi then your burstable CPU won't be as high as a node with light loads on it.
Now if no one else is using the node (e.g. Azure) then it will be blazing fast....and crash regularly. :P
Yeah ok that was a cheap shot.
Re: (Score:3, Interesting)
And it's not even a leap year!
In case people missed it the Azure network was off for an entire day last year due to a the extra day in a leap year. Since that shows a serious flaw in testing if nothing else it's difficult to take such a platform seriously.
Re: (Score:2)
Re: (Score:1)
Why not, considering that Azure will likely cost 2-3 times what AWS costs. Shouldn't MS be able to deliver a little more performance on dumb reads & writes. Even if the service is hosted on Windows boxes, surely they use something likely NAS for the actual backing store -- and the service hosting the service should be only a small portion of the end-to-end stack.
Re: (Score:1)
If you've never watched a good, smaller company go under,
Good small companies do not go under (per definition). Shitty small companies with crap products and managers to match go under. If what you are blabbing about was true, there would be no companies making software for Windows.
Re: (Score:1)
Spyglass didn't go bankrupt. They were bought in a $2.5 billion stock swap by OpenTV.
Re: (Score:2)
You must be kidding. These days Ubuntu gets more irrational hate on slashdot than Microsoft does. According to slashdot Mark Shuttleworth is killing kittens, kicking dogs, insulting disabled veterans, sniffing peoples underwear, and channeling Steve Jobs ghost.
Re: (Score:2)
Ironically, Ubuntu runs on Azure [windowsazure.com] these days (it's listed as one of the stock options when you create a new VM).
Of course it is! (Score:5, Funny)
When no one's using the platform of course its going to be faster than the competition when shared resources are concerned.
Re:Of course it is! (Score:5, Funny)
Re: (Score:2)
Re: (Score:1)
Let's see... there is nobody using Azure? So there are no iPads and no iPhones in use then? Nobody uses the iCloud? Really?
Oh, what you are saying is that you didn't know Apple didn't have the infrastructure or operating system to build a cloud offering and they therefore had to build the whole thing on Azure (striped on EC2)?
I'd say, with about 100 million iPads out there and almost three times as many iPhones sold, Azure might be the most used cloud infrastructure of all if you discount the email-only clo
Re: (Score:2, Informative)
Windows Only? (Score:3, Insightful)
>Azure is strictly windows and strictly Microsoft
Um - no. It used to be windows and MS stack only. It's not any more.
Re: (Score:1)
We are talking far more than 100 million users here.
Sure, by making things up. The Fortune 500 companies combined do not employee even half of the 100+ million people you claim.
Re: (Score:2)
3) If you happen to work for microsoft (or own shares in microsoft), please note that Azure is strictly windows and strictly Microsoft. This website is more linux oriented, and a LARGE portion of servers out there are NOT Windows based (I don't have the exact numbers, but i dare say that Linux, BSD, etc have the dominant position in the market.
I do work for Microsoft. As it happens, some of the code that I write has to run Linux, and to test it, I run Linux on Azure VMs. The ability to do so has been there since June last year, if I remember correctly.
How do they define cloud? (Score:5, Interesting)
I wish there was a standard definition of "cloud". From the article, it sounds like they are just creating and deleting AWS S3 objects and timing the transfer rates. I guess that's in line with the traditional definition of a cloud server, but I'd be more interested in seeing how EC2 stacks up against the competition.
I'm sure S3 object benchmarks mean something to someone, but seems to be an awfully simplistic measurement. It'd be more useful to see how well each service scales across many users since individual object manipulation time is only a small part of the story for people that depend on the cloud for scalability. It could take one provider more time to manipulate a single object, but still be faster when serving that object to a million users.
Re: (Score:1)
/kloud/ : Noun : A visible mass of condensed water vapor floating in the atmosphere, typically high above the ground.
Re: (Score:1)
If you've never dealt with Amazon's confusing line-up of services, then you might overlook how hard it is to compare two services to each other.
Re: (Score:2)
If you've never dealt with Amazon's confusing line-up of services, then you might overlook how hard it is to compare two services to each other.
I completely agree, but it seems a little misleading to say "Azure beats Amazon's Cloud!" when they are just looking at one tiny piece of their cloud offerings.
Re: (Score:3)
Since Nasuni is a storage vendor, it makes sense that they are benchmarking cloud storage rather than cloud compute.
Re: (Score:2)
There are a lot of definitions of cloud, true, with one of the most prevalent bring whatever will separate you from your wallet. However the NIST definition is fairly well accepted.
Not sure if test is a good one (Score:3, Insightful)
Of course it did (Score:1)
Re: (Score:1)
Azure has nothing running on it
So 100 million iPad users and maybe twice as many iPhone users is nothing? The iCloud runs on Azure.
Re: (Score:1)
You seem to be flaming this a lot. Let me clear the air for you.
1) 100 million ipad users is NOTHING. A large portion of both fortune 500 companies as well as non fortune 500 companies utilize Amazon Services in some form or fashion. We are talking far more than 100 million users here.
2) Apple utilizes BOTH Amazon AND Azure for the storage of files (and files only).
3) If you happen to work for microsoft (or own shares in microsoft), please note that Azure is strictly windows and strictly Microsoft. Th
Re:Of course it did (Score:5, Informative)
I neither work for nor own shares of Microsoft. But I have used both Azure and Amazon services. You may or may not like Microsoft as a company (or even Azure as a platform), but you're simply spreading misinformation. Azure has come a LONG way in the past year or two, and in a good way that embraces both traditional Microsoft and open-source technologies. This choice is a VERY good thing.
Re: (Score:2)
1) 100 million ipad users is NOTHING. A large portion of both fortune 500 companies as well as non fortune 500 companies utilize Amazon Services in some form or fashion. We are talking far more than 100 million users here.
There are currently around 145 million people employed in the US. The Fortune 500 in 2006 employed around 25 million people and even at an extremely high 5% growth YOY that would still only be 35 million employees. A far cry from the 100+ million you claim above and here. So, sorry, but you can keep repeating this faulty claim but its still wrong.
Re: (Score:2)
If you happen to work for microsoft (or own shares in microsoft), please note that Azure is strictly windows and strictly Microsoft
Only for the ignorant bashers. As with any religion, a little knowledge is the cure for your ignorance. I recommend you acquire some.
Re: (Score:1)
Honestly, my tongue was in my cheek, both because I hadn't refreshed lately on Azure vs. AWS usage, and because I assumed any performance study would isolate external usage as a variable. But it does appear Azure is still much less used than AWS, especially when you combine the "EC2" and "Amazon services" responses (though I'm impressed that Azure has come as far as it has in j
Yes, but ... Microsoft (Score:5, Insightful)
1. I don't know who these Nasuni people are and am not sure why I should care.
2. My experience of Microsoft over 2 decades is of a company that seems to be mostly stumbling along and not quite getting things right, occasionally hitting the target seemingly by accident, engaging in questionable business practices to ensure their market position and generally being a company many avoid doing business with if they can.
3. Amazon, by contrast has been a pleasure to do business with from my first day as a customer, continually improves it's services in a way that most people seem happy with, occasionally surprises me with things like AutoRip and is a company that most of my family (extended too) is happy to do business with. Dealing with Amazon as a partner may leave something to be desired, but dealing with Amazon as a customer is almost always a pleasure.
Given that MIcrosoft is a company that thrives by forcing customers to accept what *it* wants, while Amazon seems to be a company that thrives on giving customers what they want, my initial reaction to this report is:
Meh. So what? It's a little bit faster now - but ... Microsoft. Amazon will probably narrow and exceed that performance gap in a reasonable time frame if that's what their customers find important. So jumping over to Azure based on this report might make short-term sense, but doesn't make sense to the long-term thinker.
Addendum: What all of the above really says is that any genuine talent or ability Microsoft may have is undermined by how it chooses to do business and the resulting reputation it has earned. Which is a damned shame, particularly for all the people that work there and believe in what they do.
Re: (Score:2)
I have a history disliking Microsoft so I shall continue to do so.
FTFY.
Wait, did you, AC, just "quote" a statement you totally made up? I don't see that statement in my post, and it's not an accurate summary either. Nice try though.
There's no "history" of disliking Microsoft - I'm as ambivalent about the company emotionally as I am about any company - they aren't people, whatever the legal fictions say. I use their products or I don't, depending on what I need and the general experience I get. What I said was that I have a long experience with the company, their products
Re: (Score:1)
Wait, did you, AC, just "quote" a statement you totally made up? I don't see that statement in my post, and it's not an accurate summary either. Nice try though.
Different AC here: No, this isn't what he did. He just quickly summed up what you said in a more honest format. You seem to forget that anyone can click on your "leet" name and see your comment history. My favorite tidbit from just he first page: "Proprietary codebases are often a mess compared to a high quality open codebase."
I can understand that some people are just different. Some men swear allegiance to FOSS just as some men avoid sex with women (homosexuals) or consuming meat (vegetarians). Hey,
Re: (Score:2)
Hush and let the grownups talk.
Re: (Score:2)
Hello again, "Different AC".
Would you or "Original AC" care to comment on this?
http://hardware.slashdot.org/story/13/02/23/1415231/certificate-expiry-leads-to-total-outage-for-microsoft-azure-secured-storage [slashdot.org]
Despite your implied accusation that I just have an 'innate' dislike for Microsoft which makes my skepticism a given, thus questionable, I would in fact argue that my skepticism is well founded. The story above doesn't speak well for the customer experience for Azure users, even given the outages experi
Re: (Score:2)
But people using AWS are not Amazon's partners - they are Amazon's customers.
(ditto for MS and Azure)
Re: (Score:2)
And this is all if you assume that their findings are both truthful and meaningful in practice. They're measuring abstracted synthetics, and we have no idea if this game wasn't rigged (as is almost always the case with Microsoft products and performance).
leap day (Score:3)
Last year, Amazon Web Services' cloud came out on top, but this year Microsoft Azure outperformed AWS in performance and reliability measures.
Well, the difference is last year there was a leap day, which took Azure down for half the day, and this year there wasn't!
Microsoft has a history of forgetting (Score:2)
Like in this case [slashdot.org] a decade ago.