RSS Feed

* The Business of Software

Hi, folks.

If you're at PDC, you already know about this.  The new cloud services O.S. is entering CTP this week. 

This new way of deploying apps has a wrinkle: you only pay for the O.S. services that you consume.  No need for large up front costs to get customer #1.

This is so new, I don't know what else to say at this point, it's just exciting stuff so I thought I'd put it out. 

http://www.microsoft.com/azure

Cheers!
Greg Oliver [MSFT]

Great news!

But I can't sign in yet. There is only "Coming soon" on the page.

I'm using Amazon S3 for a year. It's really happy to get the cloud news from MS. MS is always kind for developers.

What is Windows Azure?  What is PDC?  What is CTP?

English please!

Ooh, I'll use MS cloud OS for the same reason I use other MS products!!!  But wait, I don't use any MS products at all....  :( 

Hurray for Vaporware!

The cloud arena is getting pretty competitive. Anything you think we should look out for in comparing it to others?

PDC - Programmer Developer Conference - Microsoft does this about every two years and typically they unleash a lot of new tech, some as

CTP - Community Technology Preview - somewhere between an alpha and yes we are going to do it this way beta.

I'm still waiting for The Killer Web App <echo> <echo> <echo>

So far Google is the only killer app (or suite) on the web.

An awful lot of money and energy is getting tossed at the web, but it continues to be less than inspiring.

Clouds are right up there.  2 years from now no one will use the term any more except for some old farts.  Scalability has always been with us.

Yawwwwwwwwn

WOW, who cares?!

With every year I realize I should NOT put money into Microsoft's stock.  Innovation is dead.  They've been chasing after what ever the media tells them is hot, rather than inventing what is hot.

I am old enough to remember the rise of Microsoft.  And it is pretty sad to see how they have ended up.

So it looks like Microsoft have stolen my webify idea!

How is this different to what Amazon is offering with S3 or Google is with their App Engine? It's all based on this fallacy of cloud computing as well. It's a nice idea, but limited by reality. With desktop apps your processing power scales up with your customers, with web apps it scales down with your customers.

Microsoft has become a parody of themselves, it seems. They ultimately want businesses to pay them for access to applications on a subscription basis - utility computing at its finest.

I wonder to what extent Microsoft's investment in their web technology platforms (ASP.NET AJAX, Silverlight, etc.) are geared toward this ultimate end.

Ray Ozzie says that now "You now can rely on Microsoft's data centers to host, scale, and manage your applications."

Sure. I mean, what could go wrong with that?

So... they finally released Hailstorm then?

Why does Microsoft look more and more like the IBM of yesterday?

Is Microsoft truly the place you go if you can't get a job at insurance company's IT center now?  Or the place where old programmers go to die?

"With desktop apps your processing power scales up with your customers, with web apps it scales down with your customers."

+10!

I think this is something people are not groking.  The Internet is great for what it was originally designed for.  Communication.

People have been slamming it hard with the sledgehammers of Java, ASP/.NET, PHP, etc. for years now, and today the only killer app remains advertising.

Here are two of the more interesting links on the site:

http://www.microsoft.com/azure/isv.mspx

http://www.microsoft.com/azure/howdoesitwork.mspx

Okay, so Azure is the future of the Internet?  I can't tell because in 3 browsers (FF3, Chrome and IE7!) the website looks like the worst constructed piece of crap on the entire WWW.

If Azure is the future, I want to go back to CompuServe bulletin boards.

I am bit late and trying to make sense of it.

But Azure to me doesn't appear anything special then Microsoft's Hosting Platform with their services(SQL,Sharepoint etc)!

I thought Microsoft is a software company, slowly it's trying to transform itself into a services Company. Don't do it!

If I have a killer app, why would I publish them to Azure and not my own Web server and outsource critical bandwidth hogs to Amazon EC?

Okay, now that I try following the links again in IE7 everything lines up, and it also displays correctly in Chrome. Maybe all of the javascript was making the browsers choke. Still it's not a great first impression.

So, what's new here?  Sometime soon I'll be able to use PHP instead of .NET?  Hey, that's really revolutionary.  Could someone please tell me what can be done with Azure that can't be done right now without Azure?  I have read the overview and the pages for ISVs and corporate developers and what not, and even watched the video presentation, but I still don't get it.

http://www.betanews.com/article/Windows_Azure_is_Microsofts_cloudbased_hosting_service/1225123288
Microsoft Chief Software Architect Ray Ozzie began his first day keynote speech at PDC 2008 [...] "The Web has become a key demand generation mechanism, [...] becoming Web services' front door."

"and close to 15 minutes into the speech, he was still leading up -- not yet having revealed just what it is that he's talking about"

Wow. With stimulating and technical detail like that, I now finally understand why the world is divided into "people currently at PDC" and "pond scum that will always be lagging behind MSFT's bright new future." And here I always thought it was just a big hype-fest that people pay to get into.

In any case, based on the "Live Services/.NET Services/SQL Services/Sharepoint Services/Microsoft Dynamic CRM Services" slide pictured in that article, I'm not overly excited, on a personal level. And the fact that it's apparantly for .NET languages only has yet to get me out dancing in the streets.


> Innovation is dead.  They've been chasing after what ever the media tells them is hot, rather than inventing what is hot.

Right. Next you'll claim that Bill didn't personally invent the internet.  :)

You don't invest in IBM or MSFT for abstract "innovation". You invest because you see that they have an offering that will appeal to multi-billion dollar corporations full of VPs who think that sharepoint is the most wonderful thing on the planet. Azure is not aimed at us. It's aimed at 500 person teams with a firm commitment to the belief that noone ever got fired for buying Microsoft's software running on IBM's hardware.

You invest your finances and your technical education on different criteria.


> Ray Ozzie says that now "You now can rely on Microsoft's data centers to host, scale, and manage your applications."
> Sure. I mean, what could go wrong with that?

The same thing that could go wrong if you used Amazon's EC2 service, or build your own datacenter, or even just stick a server in a closet - you could forget to make and test your backups and you could forget to have a fallback plan in case your vendor of choice sells out to a bunch of idiots and the service quality becomes unacceptable.

If the complete destruction of every datacenter on the planet used to host your application is more than a minor irritant, you deserve what you get for not having a backup plan - regardless of whether or not you think Microsoft sucks.

I'm extremely skeptical of the whole cloud thing. It seems to me that Microsoft has fallen into the trap of chasing whatever the latest Silicon Valley buzz is. I just don't get it.

@Cookie: I'd say Facebook is pretty close to another killer-app.  I was skeptical at first, but when you have a network of contacts/friends who use it actively, it's *extremely* powerful.  Microsoft were smart to buy into it early.

Of course, Facebook also understand that there's power on the desktop.  They have a desktop API, iPhone & Blackberry clients, etc.  Last.FM get that too and have cross platform desktop clients.

As for MS, I stopped using Windows as my primary OS a couple of years ago.  DRM was the last straw, Vista's bloat was the nail in the coffin.

I have to say I tend to agree with Nick Hebb.

Though I'm curious - but getting a 500 Internal Server Error right now, so I'm assuming the cloud might have floated into a tropical depression?

I wondered what they were up to with all those data centres.

It is hard to know whether this will be any good yet. Amazon's S3 has years head-start, but it is good to have a choice.

I was looking at the data service which is great, but I hope you can host a whole application on Azure, not just use it as a bit-box.

What an oddly hostile crowd.

Sure, it's not innovative, Amazon got there first. So what? I like AWS, but it won't stop me looking at Azure.

Google wasn't the first search engine, Windows the first OS etc. We routinely advise budding micro ISV'ers not to be put off by competition, why can Microsoft not follow the same advice?

Microsoft bashing was trendy like 5 or 10 years ago. I'm not a dedicated follower of fashion, is it really *still* hip?

>What an oddly hostile crowd.

Well, I have been around here for many years, and I would say that it's not odd.
On BoS it's normal to do Microsoft bashing.
I feel sorry for them, they seem to think that bashing MS is something that will make them unique.
But their arguments are, as almost always, without substance and a waste of time.

I prefer to put my time and energy where it counts, on my business.

Nope, microsoft bashing is no longer hip.

Crying about "microsoft bashing" instead of pondering the validity of other people's viewpoint, however, is definately the new black.

But it's ok, I'm reasonably sure that Bill's bank account consoles him for all the emotional harm suffered because of people posting mean things on the internet.

Also, given the number of posters here who have absolutely no clue whatsoever about the differences between Amazon's EC2 and S3 services, I'm not sure their love of or hate for Azure is entirely relevant, anyway.

@Anonymous coward:

"Crying about "microsoft bashing" instead of pondering the validity of other people's viewpoint, however, is definately the new black."

This is the BUSINESS of SOFTWARE forum. It strikes me as odd that so many people are insanely hostile to the most successful SOFTWARE BUSINESS in history.

"Anonymous posts are quite likely to be removed by the moderators"

Would that were so, my dear "". The SNR would triple.

This is the business of software forum, not the business of software temple, and therefore god does not command us to worship all businesses.

And we can even point out things they do well and things they do badly, if we're feeling adventurous.

But when you can't praise MS without being a "fanboi" and can't criticise them without being a "basher", the SNR is pretty much doomed no matter what posts get removed.

Disclaimer: I work for Microsoft, blah, blah, blah ...

I want to compare Amazon's, Google's, and Microsoft's cloud offerings, because they are all interesting, in my opinion.

I'm not going to argue the merits of cloud computing here ... that's been discussed elsewhere on this forum. It's easier to compare cloud offerings than to discuss whether the cloud is good or bad, because the latter involves predicting the future. Ultimately, time will tell whether taking advantage of cloud computing is good business for the people on the forum.

(What is it?)

What distinguishes cloud computing from traditional commodity web hosting is that it is designed to scale out rapidly and linearly. You can get started with an Amazon solution for in the ballpark of $100/month, which is actually a bit of a premium over comparable horsepower at a hosting facility. But if your app takes off, you can grow to a 10 or 100 or 1,000 server solution quickly, easily, and predictably. And if that happens, presumably cost is offset by income.

1. Amazon Web Services

So ... first of all, I was happy to see Ray Ozzie give a "tip of the hat" to Amazon in his keynote yesterday. They are the leaders in this space, and they deserve kudos. And of course, it's easier to give them kudos now that they support Windows in their cloud.

A cloud platform involves lots of capabilities, but at a minimum, you need storage and compute (authentication is handy, too, and there are lots of other "nice to haves"). Amazon has a couple of different kinds of storage. S3 is a "bit bucket" solution. Very easy to store and retrieve a stream of bits in a named location. SimpleDB adds querying capabilities (and was built on S3 and EC2). They also have block storage, which allows you to mount something that looks like a disk, so you can run big honkin' instances of Oracle and SQL Server and the like.

The compute side of Amazon is the Elastic Cloud or EC2, which is essentially a platform for managing virtual machines. This is important - the COMPONENT in EC2 is a virtual machine. Typically, this vm is running a stateless web stack (e.g., Linux/apache/php - but no database). In order to be able to scale out, you can't store state on these virtual machines - one vm instance has to look just like another. They do DNS tricks to balance load across VMs.

So Amazon's compute is based on a "VM-as-component" model.

2. Google App Engine

Google's App Engine takes a different approach. I think of it as the "sandbox" model. Their cloud platform currently only supports Python (I suspect Ruby will be next), and a component is a chunk of Python code (typically, a web page, I think). Again, in order to scale out, the compute "chunks" have to be inherently stateless, so the App Engine "sandbox" disables features like accessing the file system. All state has to be managed through a specific data API.

What's intriguing about Google's offering is that it's pretty much the way they deliver everything themselves. It's easy to imagine that as they improve their own apps and services, the App Engine will improve right along with them. So if you love Google's apps and you love Python, well ... it's clearly worth a look.

3. Microsoft Windows Azure

Yesterday, Microsoft unveiled Windows Azure. It's very early (preview bits are currently only available to the people who are attending this week's Professional Developer Conference, but I apologize for the lame "coming soon" explanation on the site ... should have been clearer).

Microsoft's vision is arguably the most ambitious. It's a sandbox model, not a VM model (which makes sense - after all, you can already do Windows VMs at Amazon now, so that choice is already available to Windows developers). At first, only .NET "components" will be supported, although native Windows code support is being targeted for future releases. Storage includes SQL Services, which as you can probably guess is a queryable data store that can scale out enormously (and will eventually include reporting services, analysis services, and other features of SQL Server, since it is built on SQL Server).

The biggest advantage of Microsoft's cloud for the readers of this list will be the development experience. Soon (hopefully a couple of weeks for readers here), you'll be able to open Visual Studio, create a new "cloud app", and be running a web page or web service in the cloud - with storage in the cloud - pretty instantaneously. You'll be able to run it locally for test and debug using tools you're familiar with if you already do .NET. And it will be priced "competitively" ... I have no idea what that means, but we have a good track record of having low prices.

Windows Azure will be useful for general purpose web applications ... and in fact ideal for creating pure Saas applications (or better yet, the back end of "Software plus Services" applications) for enterprises and consumers. Then there are the Live Services side of Microsoft's cloud offering, which are pretty focused on consumers. It's a complementary model that requires an even bigger shift in mindset for MicroISVs. Instead of writing web pages and services (which Windows Azure will allow you to do), you'll be creating functionality "for the mesh". Data and apps sync themselves across all your devices. Users like being able to run an app on a PC for the richest experience, on a phone for "spontaneous" usage, and via a web browser for those times when they are away from their machine. Live mesh and the rest of "Live services" will make that type of experience realistic and perhaps even enjoyable for MicroISVs and startups to implement.

Summary:

AWS => Most mature offering. VM model (Linux and Windows), which is familiar and easy but requires the developer to think about scale a bit differently (i.e., they might screw up their first try).

Google App Engine => Just like Google. Sandbox model running Python.

Microsoft Windows Azure => Familiar to .NET developers. Sandbox model plus ambitous new services. Early preview bits coming online over the next couple of weeks.

Hope this comparison was useful to somebody. Much more to talk about over the coming weeks and months.

I agree with the blank poster above. Being critical of a company's actions does not make you a basher. Personally, I like Microsoft. I just think they are heading in the wrong direction on many fronts.

@Tim, I have been on the Microsoft Platform all my development career spanning 10 years. But I am critical of Azure (I think pronounced "Assure" with a z?). The thing is it's not really an OS on the internet as it is touted. It is infact just another hosting platform and there is nothing new to this except for Marketing Hoopla surrounding it.