Hacker News new | comments | show | ask | jobs | submit login
Microsoft Azure Sales Top $1 Billion, Challenging Amazon (bloomberg.com)
100 points by mandeepj 1516 days ago | hide | past | web | 98 comments | favorite

Love this quote:

“The biggest advantage I have with Microsoft is I don’t have to go to any other vendor for any solution -- I can go to one partner for all of my operating systems, all of my development environment and all of my infrastructure tools,” he said. “Why would I waste time looking at another third-party solution?”


"Nobody ever got fired for buying IBM". "I'm a lazy C-level exec." "I love having all my eggs in one basket." "I just write a check, and all my problems are solved." "Money buys happiness."

The flip side is that you end up in a scenario where two or three vendors can blame each other for problems. With Azure, if SQL has a problem, Microsoft has to fix it, and can't blame any other vendors, driver, hardware, or network issues.

Edit: Also, really? You don't think there's any benefit to integration? That's like Microsoft's whole deal - make their stuff work well with their other stuff. Sure, it creates lock-in, but it's also a feature users want. I want my various programs to easily work with each other in obvious ways, and not have to worry about compatibility issues.

> With Azure, if SQL has a problem, Microsoft has to fix it,

Not really. Vendor lock in still applies.

>Not really.

Mind explaining a bit more? If I buy a Samsung Galaxy S4 and the lock screen wouldn't let me in because of a bug in Android, Samsung is still responsible.

>Vendor lock in still applies.

Again, how exactly would that be, if I am say using Ubuntu and Mysql+PHP on a Azure?

And how much effort is Microsoft willing to dedicate to fix your Ubuntu/MySQL/PHP stack?

It's only a single vendor solution until you introduce a second one.

>And how much effort is Microsoft willing to dedicate to fix your Ubuntu/MySQL/PHP stack?

Isn't that the same as asking how much Google is willing to fix Linux for the Android stack??

> It's only a single vendor solution until you introduce a second one.

And your solution is to tell everyone to get a VM and install and maintain everything from scratch? Yeah that will definetly work, which is why the internet is secure with up-to-date Ubuntu stacks from all the daily "sudo apt-get update" keystrokes. /s

A lot actually - the node.js solution on Windows Azure is super solid and they have a core developer on the node.js project to make it so.

I see your points but I think your characterization of this attitude as lazy or stupid are overly hostile.

I'm not going to defend the mindset but I can see some appeal in having one company to deal with. Anyone who's ever faced a problem where multiple vendors are blaming each other for your ongoing woes knows how frustrating that can be.

And that's what a CIO/CTO's job is. Be technically competent to understand the technology he is using, or (at least) have someone on his team that is. It isn't rocket science to figure out where the problem is (be it CPU, network I/O, some software run wild), etc. If you're involved in IT, especially if you're developing software, you should be able to do some rudimentary analysis of the software you're actually buying.

That's a very simplistic view. As many AWS, GAE, and Heroku clients will tell you, when your app is up there in the cloud, there are times when all the instrumentation in the world won't help you fix a problem.

C#, .Net ASP, Visual Studio, and Azure are all very good development products. I know, because I tried them myself. Or ask the Stack Overflow guys. It's called thinking for yourself. As for your appeal to the recieved wisdom that everything MS does must suck because Graham said so, let me point out that their monopoly gives them very little advantage when it comes to catering to developers. No one has to use their servers, languages, DBs, etc. So they have to do a good job.

That's an odd quote, especially when most competitors provide windows virtual machines, and azure provides linux. Also, I doubt msft cancels your account if you use mysql instead of sql server.

...but I honestly agree with the sentiment. I will pay money to not deal with sysadmin headaches, which is something you get a lot of using a hodgepodge of technologies.

There are a lot of people building on top of MSFT platforms. ASP.NET, C#, etc. may not be sexy, but a lot of people use these languages/platforms regularly. Probably more than any particular web-based language. Being able to your desktop skills and move them to the "cloud" with tight integration in Visual Studio helps leverage a large existing developer ecosystem.

Comparing against Amazon's total cloud sales may be comparing apples and orange. To quote from the article:

"Microsoft’s $1 billion sales figure includes Azure, as well as software provided to partners to create related Windows cloud services, Anderson said in an interview."

What software is included "to create related Windows cloud services", and is that the only use for said softare? i.e., are they including things like MS Visual Studio? And how much of MSDN subscription fees included in the $1 billion dollar figure, I wonder?

I wonder if the $1B number doesn't also include internal marketing "funny money" microsoft is using to seed the business. There are companies who have been given large stipends of free services to wet their appetite for the platform.

Most likely the software sales are Windows Server licenses and the Azure tools needed to make a private cloud. A lot of companies have policies that prevent them from using public cloud servers.

The numbers are fishy, but if they are growing by XX% (1) a year, expect MS to reach that pretty soon. Love it or hate it, getting all your tech needs from one company has its advantages (and disadvantages).

1. Azure subscriptions have risen 48 percent in the past six months, said Takeshi Numoto, Microsoft’s vice president for marketing for the server and tools division. That unit, which encompasses Azure, has posted nine straight quarters of sales growth of at least 10 percent, he said.

The server and tools division is a big business and the fact that it's revenue is growing at 10+% for 9 quarters is impressive.

It's easy to rise 46% in terms of subscriptions when they offer tens of thousands of dollars worth of their service for free.

Amazon micro instances are basically the same thing...

How are the free hours from BizSpark handled? (http://www.windowsazure.com/en-us/offers/ms-azr-0012p) Are those treated as sales?

Even if they did, they probably don't add to the 1B figure as they are free

I don't want to describe this comment as ignorant but I almost feel that I have to...

There are numerous accounting tricks employed by organizations big and small to book revenue on promotional services.

One popular way of doing this is to charge the marketing departments budget for the promo hours but record the cost of the hours as revenue. The overall bottom line of $MSFT is uneffected but the product, in this case azure, benefits from increased visibility.

I don't think Azure is doing anywhere near AWS in sales if you don't count promo deals.

Completely agree with this. They are giving away massive amounts of platform time to seed usage. Better question is when Azure will break even (net, including R&D and internal marketing transfers to the Azure business).

"“The biggest advantage I have with Microsoft is I don’t have to go to any other vendor for any solution -- I can go to one partner for all of my operating systems, all of my development environment and all of my infrastructure tools,” he said. “Why would I waste time looking at another third-party."

^Interesting viewpoint.

I'm an MS dev with a little Django, Rails and PHP experience and I'm always sort of overwhelmed with the choices I have non-ms environments.

Recently I wanted to get a simple laptop for Python development and I don't know how to begin selecting the right tools for the job. I see myself having to make all sorts of evaluations I'm just too inexperienced to make. Which OS? Whats PyPy good for? Is it supported on Ubuntu? Oh I can use this command to install but I don't see an apt get example? Does that mean it won't work?

When I read about and fire up EC2 instances I feel the same way. How do I configure them? Which one is best, etc.

With Visual Studio and the Azure tooling built in a lot of decisions are just invisible to me and I sort of like it that way.

Of course I understand the benefits of open source and having a variety of options is important but I'm ok with giving up a little control and power to just get things done.

I think its the same reason we use Stripe instead of getting out own merchant account and all that rigmarole.

Isn't Heroku similar as in Heroku makes some choices for you on how to do things?

Knowing how to find the answers to such question is an element of being an expert - not knowing all the answers, but knowing how to find them.

Djanjo, Python, Rails, and PHP can run find in a Windows environment, but the community is rooted Linux (and OSX to some degree), thus the amount of quality resources available for your research is greater. Same goes for .NET - more Windows information is available compared to Mono.

Being able to make the best choices for a project is a valuable skill.

Totally agree with you however what I'm finding to be the most difficult is a lot of getting started resources tend to be using different nuances of the Python programming environment.

My goal is to setup an environment closest to what I'd see in the field but its really hard to determine what that means.

In the ms world when you bounce jobs getting up and running usually involves using the same three programs you always use.

From reading about python I'm getting the sense that there is so much more variety out there in the field.

This comes down to what your goals are.

If you want to do Django, just use this:


It might not be the latest, the best web server, etc. but it will get you going.. Give it a few weeks and you'll be better able to answer your own questions.

At this point, starting is the best approach.

I'd suggest learning Ubuntu - it's popular and widely used as both a desktop and server OS, and it's great for running Python.

Run Ubuntu inside a virtual machine (VMware or VirtualBox) and you can simulate your exact deployment environment on your local laptop.

You know, to some extent, this is a direct result of working with Microsoft tools. I am generalizing but Microsoft developers get to used to getting mostly nice and tidy answers to everything (regardless if its good).

With open-source, my gosh, you actually have to do some analysis and make a sound judgement without having the Microsoft mothership tell you whether its good or bad.

I don't dispute that the choices can be seen as overwhelming but this is where experience and your own curiosity work best. You don't have the experience, well there is many many people that do and are willing to share their knowledge. It does require a desire to learn. The more you investigate, the better your filter on what is good or what is bull improves.

I use Microsoft tools, they have their place. It's not a question off giving up control and power, its a question of, what makes you better? Understanding the dirty work that it takes to make software, OS, web frameworks, etc. work, or having a nice little button that does this for you (I know this is a gross exaggeration).

If you have that curiosity to understand the details even behind Microsoft tech, using Microsoft is just another tool.

There's a lot of development and apps that simply are not going to have any real difference based on their choice of ORM or even DB. Having it all-in-one has a lot of benefits.

Also, it's unfair to say open source isn't the same. How many PHP projects fairly evaluated anything but MySQL or PHP's built-in MySQL data access libs? And why would you? If you aren't at the point where you need to evaluate, your app will probably work fine on MySQL anyways. Also, Ruby-on-Rails provides quite a bit of defaults...

Microsoft might provide more kinds of libraries, but that's only because of their size. If Zend created an MVC system and built it into PHP, what kind of uptake would that see?

"With open-source, my gosh, you actually have to do some analysis and make a sound judgement without having the Microsoft mothership tell you whether its good or bad."

I think you're comment is fair but a little condescending.

Speaking for myself, I play and experiment with tons of non-ms libraries for asp.net work, ran Ubuntu as a desktop for several years, have an trivial app on heroku.

Just because somebody appreciates the simple things does not mean they don't have curiosity and don't appreciate knowing technology on a lower level.

The big divide between the ms and non-ms communities is because of this sort of comment.

Not meant to come across that way (especially towards your comments).

I definitely agree that there is value with a full life-cycle integrated approach that Microsoft offers with their stack. I really like what Microsoft is doing with ASP.NET MVC. I also like their move towards embracing the best of open source such as their recent full support of git within their stack. I setup a gitlab virtual machine and using the Visual Studio git extension with it has its quirks but works really well. I find this a really good compromise of mixing the best of both worlds.

I guess my main angst is with Microsoft stack developers who I've worked with that really don't want to step outside their comfort zone. It's definitely obvious you are not within this group.

Well, Visual Studio probably has the best Python plugin I've ever seen.

What most people on HN miss (being the developers we are) is that having one relationship with a vendor is hugely preferable for non-development reasons. Billing (simpler, also savings in bundling), unified support (why doesn't x work with y!) and so on mean that going to a vendor isn't just the lazy choice it looks like.

It's only interesting until the need arises to go elsewhere. Then it becomes really bad, since MS doesn't make it easy after you are too used to not paying attention to the pitfalls of lock-in.

Microsoft's wet dream customer.

Yes but an amazingly common one. If I look at the NZ city I live in (pop ~80k) both Tertiary institutes, City Council, Regional Council, primary Health organization etc are all very much happy Microsoft shops.

They say that because it isn't their money they are spending.

NZ seems to have a pretty large and active MS developer scene.

Having worked in India for more than 3 years, most developers I knew hadn't ventured outside of the MS ecosystem. Xerox's business services divisions (who made this statement) likely satisfy large enterprises working on MS technologies for Customer Care, HR, Payroll using the readily available workforce in India.

How much of this is influenced my vendor participation in schools?

I experienced a situation in a US university where a computational statistics program was more about SAS than the theory in general. The program was more about a brand certification, than an overview of tools and techniques.

IMO, the approach in a public university is poisonous.

A viewpoint that fails the "right tool for the job" test. Sometimes it is a Microsoft platform, sometimes not.

There is a major segment of tech that build careers on brand, not skill, and thus the quality of their solutions equally bottlenecked.

Are Apple iCloud services still hosted with Azure tech? http://apple.slashdot.org/story/11/09/04/0051209/apples-iclo...

Because that alone could be a large portion of that $1B

Yes, as far as I know Apple hosts their iCloud jointly on AWS and Azure.

Which is a great strategy for disaster recovery or failover I think, considering both the cloud services are the largest and are owned by financially stable companies that have been around for a while.

Who here on HN is running any portion of their startup or established Internet business on Azure?

I'm hosting some (really) non-critical stuff on a Ubuntu image on it.

It seems fine. The manager interface is actually pretty good IMHO, but Linode is really cheap and running Linux on Azure is (I think) still a "beta" feature. I get the hours free though BizSpark. I was thinking of setting it up as a slave/mirror of our main sites for use as a fallover backup. It's been pretty reliable but I haven't done any benchmarks or really tested it in any way.

If I had need or inclination to use the Microsoft stack, though, it would definitely be a serious contender.


Sounds like it was just because it's free. (BizSpark)

Working on a project right now that will deploy on Azure Platform (not VMs). It's pretty nice to be able offload the OS to a third party and just deploy your app bits, knowing they're going to take care of OS, patching, and in general, server management tasks.

On the VM side, they're still catching up, I think. For instance, you can't get multiple IPs routing to a single VM, so if you have multiple SSL sites, you'll need multiple servers in general. (SNI isn't available on IE for XP which is a deal breaker for many people.)

But that's probably a reflection more of the mindset of people wanting dedicated servers and thinking "cloud VM" is the same thing, instead of people trying to build distributed applications.

I've worked with/for 3 companies all built completely on Azure. I was hired by 2 of the 3 companies to help them switch to Azure. The third is my own startup which has been on Azure since Microsoft launched it.

It really is a fantastic platform.

I played around with Azure and Heroku for node.js hosting and found that Azure wasn't as painful to setup and didn't trip up as often as Heroku (though my statement should be read as more of a criticism of Heroku than an endorsement of Azure)

I'm running an "established Internet business" on Azure. I moved from VPSes to Azure last year when the "Azure Websites" feature was available. It has saved me money over the VPS instances, actually increased performance a bit at the lower price point, and has dramatically lower administration overhead since I'm using it as PaaS.

The only thing negative I could say is that I have a hard time logging into the portal sometimes, due to some weird interaction between my primary MS account and one of my clients' Office365 organizational accounts. So, I sometimes have to switch to an incognito window to log in and use the portal. It's a nuisance, but trivial.

Otherwise, I couldn't be happier.

I run several services on it. .net/C# compute services. No VMs. No complaints.

different question: if you aren't why not? Amazon hasn't exactly been the picture of reliability and Heroku has had some important unanswered questions raised about its dispatch system.

building a new service on Azure here, started 3 monts ago, switching completely from LAMP to .NET. I love working with VS and I'm getting to know Azure setvices better day by day, so far so good.

Please everyone, not all at once.

I like how HN is so blatantly anti-Microsoft, but Reddit is overly Microsoft supportive. Which hipsters do I believe in?!

It depends what subreddit you go onto, HN is pretty small compared to the reddit hivemind.

I don't think it matters. If Azure is challenging AWS, Azure is challenging AWS. Who the buyers are is irrelevant; HNer dollars don't double when you deposit them in your bank.

I make a quip, and all of a sudden it's something to believe in? Humor me more :D


There is absolutely, positively, certifiably no way this is even close to true. Did the people are bloomberg even look at Microsoft's figures, or were they too busy counting the cash MS sent over?

It seems pretty reasonable to me. If Msft makes $1 billion on 20% of the market, and Amzn makes $3.8 billion (high-end analyst estimate) with 71% of the market, the math roughly matches.

I think those market share numbers are suspicious.. according to rackspace they had 1.3b in sales of private cloud and public cloud services in 2012: http://ir.rackspace.com/phoenix.zhtml?c=221673&p=irol-ne...

So either the total is more than 100% market share for the 3, or those numbers or wrong.

Edit: added "in 2012" for rackspace financials

I'd agree, most of these numbers are "fuzzy" and based on analyst estimates; I'd almost be more surprised if they DID perfectly add up to 100%. :-)

The percentages aren't market share rather percentages of companies using the product. That is a company which uses both Azure and AWS could contribute towards both percentages.

I think you underestimate the penalties for lying about something like this in a publicly traded company, false statements about sales is securities fraud.

You massively underestimate enterprise/corporate use.

And especially how a good sales team can inflate revenues in that market...

It's true if you bend the rules a bit... Read the article and you'll notice this gem:

> Microsoft’s $1 billion sales figure includes Azure, as well as software provided to partners to create related Windows cloud services, Anderson said in an interview

That is a really vague statement that could mean all sorts of things (Office 365, Windows Server licensees, etc).

You do know that Microsoft is dominant in Enterprise and that that is where most IT dollars are spent.

The more surprising thing to me is that Microsoft isn't burying Amazon. But give them 2 years.

It's clearly a puff piece. Having said that, there are a lot of MS server customers and I can imagine Azure being a compelling option if you don't want to do a rewrite.

Can Microsoft's internal Azure expenses be accounted for in sales? Specifically, if Office365 used $100 million worth of Azure services in their budget, does that count toward overall Azure sales? Does part of the marketing budget go toward giving out free Azure trials? Then, the question I ask is what percentage of Azure sales come from 3rd parties.

Doesn't surprise me. Microsoft has a strong hold in the "enterprise" (read large corporations with bureaucratic but big pockets) sector.

>"They haven’t excited the front-line developers -- the ones who made Amazon who they are," he said. "Those will be hard to influence."

That may be true, but at the end of the day, startups aren't generating multi-millions/billions to spend on large server deployments.

That looks pretty significant. Amazon Web Services are somewhere between $2 and $3.8 billion (depending on which analyst you believe), so Microsoft is potentially within striking distance of taking the crown.

Could definitely be true if revenue from internal customers (Office 365, Hotmail, Xbox etc) is included.

I don't think either Hotmail or Office 365 use Azure in non-trivial amounts, or atleast they didn't in my time there until 2011.

The next time you see an article mentioning Microsoft's Online Services Division, it's history of losses, and suggesting that Ballmer should be fired, just remember that Azure's revenue is not part of the Online Services Division (and neither is Office 365).

I've been using azure a bunch at work. I don't know where the sales numbers come from. I have reasons for being skeptical after being in on conversations with their sales reps. I hate how they hide important details about their systems. The shroud of mystery is really infuriating.

Useful bits in terms of performance metrics just aren't there or just aren't visible. I've used Amazon's services and Rackspace's services as well. I've had mixed experiences with all of them. I honestly feel that I'd be getting work done much quicker if I was just handed a vmware portal and managed VMs in a datacenter to implement our systems. It's very nice when you're able to SSH into a system and run top or other tools to see what's going on. Running queries or checking out charts that show data at hourly intervals is next to useless. As an added bonus, I wouldn't have to stop queries and go up to the application layer and reconnect to another shard in order to continue my query.

I honestly don't know what's so appealing about Azure. Maybe devs these days don't know how to ping a server or set up a database. If it weren't for getting tons of free cash to start, we'd be on about 1/10th the "budget" using other services. But it's hard to beat free.

Would I ever pay for it? Yes. But only in the situation where I had a bunch of sub-par developers. Id also be looking for an out from the company while getting patted on the back for going with MS.

They are fast too, they already have ubuntu 13.04

I LOVE Visual Studio and working in C# / Asp.Net MVC. I do it full time at work, and now writing my SaaS in it which I plan to host on Azure (though, I wish they had an official MongoDB offering).

I've always done PHP on my side stuff, partly because of the hosting issues, but now with PaaS' like Azure and Appharbor, people who enjoy the MS Stack at least have reasonable options.

People bitch about MS, but without competition from Azure, Google, and others, Amazon wouldn't keep lowering their rates.

anyone know what app engine sales are? I don't recall ever seeing it mentioned in investor relations stuff makes me wonder why or if/when they will kill it.

Considering the pricing on their server software, I imagine at $1 billion revenue they must have a solid 10 happy customers.

Knowing the price of Azure that should mean they have roughly pulls out calculator and hits some keys.. 12 customers.

Microsoft's prices for Azure are actually very reasonable, and I generally find their tools easier to configure and use in real life. Microsoft makes some horrible products, but I actually really like Azure a lot.


I checked this out as a pretty heavy AWS user, and I have to admit that GUI looks pretty nice.

Make sure you benchmark the VM's if you intend to use them - performance was horrible during the preview. http://openbenchmarking.org/result/1206086-BY-1103157IV31

Aren't Azure's prices exactly the same as Amazon's?

Sometimes you only need to have 2 - if they are like Nestle or Exxon ...

Thank you for contributing such a thoughtful and valuable comment to this discussion. It really serves to improve the signal to noise ratio and raise the level of intelligent discourse here.

Kiss my ass.

This seems highly unlikely considering they don't remember to update the azure SSL certificate two years in a row.


It's not "the azure SSL certificate", it's more that there are a great many of them being used internally between many systems.

It caused a global outage for SSL access to their cloud storage. Two years in a row. Close enough.

They were two very distinct issues, both in retrospect, should have been preventable. One was a cert not getting renewed correctly (which is really more of an engg/ops process thing) and the other was a piece of code that did date checks incorrectly.

[Disclaimer - early and long time Azure guy at MSFT]

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact