Comparing Amazon's EC2, Google's App Engine and Microsoft's Azure
The weather forecast changed when Microsoft entered the clouds with the Azure platform during PDC 2008. It would be interesting to compare the three major offerings existing on the market today, Amazon's, Google's and Microsoft's, and at the first glance it seems that they are not really competing against each other.
Michael J. Miller, vice president at Ziff Brothers Investments, compared the three major players in the cloud computing and this is what he's found about Amazon's EC2:
Clearly the cloud platform that has gotten the most attention is Amazon Web Services, which is a collection of a variety of tools, mostly at a very low level. The one that gets the most attention is called the Amazon Elastic Computer Cloud (EC2), a web service that lets you assign your application to as many "compute units" as you would like, whenever you need them. To give you an idea, a single default instance includes 1 "virtual core," 1.7 GB of memory, and 160 GB of instance storage (which is storage that is used only for that session) for 10 cents an hour. On top of this, you might want to use the company's "simple storage solution" (S3), which costs 15 cents per GB per month for the first 50 terabytes of data, and goes down from there, plus charges for certain transactions. You might also want to use the company's "Simple DB" database, or its queue for storing messages, which have additional charges.
The Amazon platform's basic advantage is simple: you can just use the amount of storage you want, when you want it.
This is what Michael says about Google's App Engine:
Google's App Engine is newer. It is still in a free beta stage, and the tools so far are a bit more restrictive. As I understand it, where Amazon gives you a virtual machine you can install pretty much any software on, Google pretty much gives you a fixed environment, based around the Python language, the Django development framework, Google's BigTable database/storage system and Google File System (GFS). For now, developers get 500 MB of storage and compute power for up to about 5 million pageviews per month for free, and the company has announced pricing for more active sites. For instance, the company says developers should expect to pay 10 to 12 cents per CPU core-hour.
Because this has been built so closely around Google's own operating environment, it should be relatively easy for developers who know those frameworks to get started. But some developers have shied away from it because it seems more restrictive than Amazon's solutions.
When it comes to Microsoft's Azure, Michael affirms:
Like Amazon Web Services, Azure actually consists of a variety of different services on top of a common platform. ... The .NET services are the ones getting the most attention now, because that is how developers would write for the platform initially. Indeed, from the sessions I attended, it looks like taking applications written for the .NET framework and developed in Visual Studio can be moved to "the cloud" relatively easily.
One big difference with Azure is that while Microsoft intends to offer its own hosted Azure services, the platform is designed to be able to run on local computers or corporate servers as well. This should make it easier to test the application, but also to allow corporate applications to be run within a company's network as well as outside.
Michael comparison conclusion is:
So looking at these three players, you see each of them playing to their strengths. Amazon was early in the market, and has leveraged Internet standards and open source platforms to create a very flexible platform. Google is leveraging the work it has done with big databases and its internal development methods to create a powerful but more restrictive environment. And Microsoft is leveraging its traditional strength with developers and the breadth of its tools to create perhaps the largest array of services. Over time, my guess is that we'll see all of them start to converge - presaged perhaps by Amazon's introduction of Windows Server instances.
Jessica Hodgson from Dow Jones Newswires wrote about the new financial equation generated by Microsoft's entering into the game. She quotes Matt Rosoff, an analyst with research firm Directions, saying:
What I think Microsoft is trying to do is to freeze the market.They want people who may have been thinking about experimenting with on-demand products to hold off and wait to see what their products are like.
There are various opinions about the value of the cloud computing market, according to Jessica:
Though everyone agrees cloud-computing is growing, opinions differ as to whether it will replace on-premise computing or simply develop alongside it. Deutsche Bank analyst Tom Ernst says the market for packaged software has already peaked. Within five years, Ernst estimates cloud-computing services will account for roughly half of the $60 billion application software market. By contrast, Oracle Corp.'s Chief Executive (ORCL) Larry Ellison, has mocked cloud computing as "gibberish", and said few companies will be able to profit from it.
Microsoft's commitment to the business model could boost expectations, prompting big companies that have been reluctant to embrace cloud computing to adopt it. Richard Campbell, co-founder of Vancouver-based Web developer Strangeloop Networks, says many of his customers are asking about cloud computing, though few have switched, largely because they have concerns about security and reliability.
Jessica continues analyzing Microsoft's move:
Microsoft is moving carefully, protecting its packaged software franchise while assessing how the cloud computing market will develop. It has adopted a hybrid approach because it says most customers will continue to want on-premise products, a position supported by many analysts. ...
If Microsoft moves too slowly, it risks letting innovators, like Google and Amazon, grab market share. Those companies don't face Microsoft's profitability dilemma because they don't have physical software franchises to protect.
"If it becomes a popular way of doing mainstream business, it's hard to see how they can avoid cannibalizing sales of their desktop services," says Nicholas Carr, a former editor of the Harvard Business Review who has written a book about cloud computing. "It's going to be a scale game."
Oracle Cloud Computing