News Blog

Looking to take on more demanding customers, Amazon Web Services on Thursday rolled out two paid-support plans that give customers access to its engineers to resolve glitches.

The company said it will offer two levels of support--gold and silver--for a fixed annual fee or a percentage of customers' total usage of its services. The support plans are available for its Elastic Compute Cloud (EC2), Simple Storage Service (S3), and Amazon Simple Queue Service (Amazon SQS). For more details on the terms, click here.

Right now, Amazon offers pay-as-you-go pricing for its hosted services. Customers pay for how much they use the service. To get support for technical problems, they need to go to free forums.

The paid support is a sign that Amazon's hosted computing is ramping up to take on a broader swath of clients, including large businesses.

Initially, Amazon aimed the hosted service at Web start-ups, but it's signing on business customers too. BusinessWeek reported earlier this week that The New York Times and Nasdaq are now customers.

The support service also casts Amazon more in the mold of traditional IT providers such as IBM, Hewlett-Packard, and Sun Microsystems, which all offer a variation on hosted computing.

"Guaranteed support will also allow us to develop even more substantial applications using Amazon Web Services, knowing that Amazon is there to support us," Paul Horvath, chief technology officer of health care form-processing company TC3 Health, said in a statement.

Update on Friday: Link added to Amazon Web Services support terms and costs.

As people get their heads around Google App Engine, they see some things they may not like. Namely, the dreaded "lock-in."

Developers for years have been clamoring for more openness and standards. They are tenets of the open-source movement.

But as more application development moves to hosted platforms, does data and application portability get lost in "the cloud"?

Given that we're at an early point in platform-as-a-service offerings, I'd say lock-in, to some degree, is inevitable. Most people consider Salesforce.com's Force.com closed, as it's based on the company's database and query language.

But Google? The search giant is hosting a Web development conference next month, not to sell more software stacks or subscriptions, but to encourage more apps--and people--to move to the Web, it says.

Still, O'Reilly takes Google to task for the lack of application portability--at least in this first iteration of Google App Engine.

"Now, it may be that this is a temporary oversight, and that Google does intend, long-term, to make it easy for developers to export their applications. After all, Eric Schmidt says he reminds his employees all the time, "Don't fight the Internet."

But it's also possible that this is one more sign that one of the big guys is forgetting the principles--the Internet as a platform (not "my company as a platform"), harnessing the power of user contribution (which, as John Musser pointed out, means that you always "pay the user first"), small pieces loosely joined--that brought their success in the first place.

Think his concerns are overblown? He's not the only one.

Within a few days of its release, programmer Chris Anderson wrote some open-source software, called AppDrop, that shows that you can conceivably run an instance written for Google App Engine on Amazon.com's Elastic Compute Cloud (EC2), Amazon's hosted server platform.

Developer Alex Bosworth listed lock-in as his top concern with Google App Engine.

It's likely that Google will allow applications written with other languages, like JavaScript. But the nub of online-platform lock-in comes from the data store, Bosworth said.

One thing both Amazon and Google could do to really show they are serious about their platforms is open up their data engines, which are really the core of most Web applications--open-source BigTable and SimpleDB. This would really reduce lock-in and make development easier, and it might even lead to some help improving their services.

O'Grady at RedMonk, too, argued that Google should open-source portions of its infrastructure or offer an API (application programming interface) to its data store that would ease portability to other databases.

Google appears to already be on the case of data portability. On the Google App Engine Blog, software engineer Kevin Gibbs said that one planned feature is large-scale data import and export.

"With Google App Engine, you own all the data in your app. As stated in our terms, you always have the right to get your data out of Google App Engine at any point. We wouldn't have it any other way," Gibbs wrote.

Once again, Google gets tongues a-waggin', even when it isn't the first to a party.

But it's good to see these issues raised and for developers to push for more openness. After all, standards, portability, and interoperability have been good to the Web.

It's just like an unformatted hard drive, Amazon.com Chief Technology Officer Werner Vogels explained. The difference is that it's in the "cloud" somewhere and you get to it through an API.

Amazon Web Services executives on Sunday described a forthcoming persistent storage feature, called EC2 Persistent Storage, which they say will make its hosted computing services more flexible and far more reliable.

People can sign up for an early beta test program now before Amazon opens it up for a wider release later this year.

The service works with Amazon's Elastic Compute Cloud (EC2) hosted server offering. It allows developers to set aside a storage volume online, on which people save files in different file systems. This differs from what is available now with EC2 because once a compute instance is taken offline, the data associated with it goes away.

With a persistent storage service, data can remain linked to a specific computing instance. Significantly, people can take a snapshot of that data and store it on Amazon's S3 storage service. That effectively acts as a way to create a back-up of their computing operation on the "cloud," according to Amazon executives.

"The snapshot is extremely powerful technology and allows for building highly fault-tolerant applications operating worldwide. Combine these snapshots with Availability Zones and Elastic IPs and you have all the tools to manage and migrate even the most complex of applications," Vogels wrote on his blog.

"And the great thing is it that it is all done with using standard technologies such that you can use this with any kind of application, middleware or any infrastructure software, whether it is legacy or brand new," he added.

Amazon Web Services evangelist Jeff Barr also describes the service on his blog, saying it was one of the most requested features from developers.

Thorsten vok Eiken at RightScale, who has been testing the service, talks about the implications of this feature and says his company is making tools to make it easier to use these services.

Von Eiken says that persistent storage is a dramatically important feature that will lead many more companies and developers to hosted development platforms.

"It's going to be like agile software development: if you want to survive as an Internet/Web service you will have to compute in the cloud or your competitors will leave you in the dust by being able to deploy faster, better, and cheaper," he said.

Amazon.com's Elastic Compute Cloud Web service was knocked offline earlier Monday, but the company appeared to get it back online within a few hours.

A thread on the Amazon Web Services support forum started at 1:51 a.m. PDT. By just past 4 a.m., a support person said all servers that had been unreachable could be contacted again.

(Credit: CNET News.com)

It's a far milder outage than the one that occurred in February, when Amazon's Simple Storage Service went down, which appeared to affect hundreds of Web sites.

It's also a reminder of the scrutiny on new hosted computing services, which many software developers are using because of the low up-front costs and simplicity.

Although there has been talk of utility computing for years from the likes of IBM and Sun Microsystems, Amazon is increasingly recognized as a leader in the field.

Google is rumored to be making an announcement on Monday of hosted database services around Bigtable, which my CNET News.com colleague Dan Farber described here. Microsoft, too, is expected to offer more of these sorts of services as well around its Live development services.

Customers of an Amazon Web Services support forum lauded Amazon for providing regular feedback on the company's efforts to resolve the problems. A representative of Amazon also said it would hold an investigation to get more details on what caused the problem.

Amazon Web Services on Thursday is scheduled to release features meant to give its hosted computing service a better safety net.

Amazon's Elastic Compute Cloud (EC2) service now has an application programming interface (API) that lets developers choose where its application physically runs.

"Up until now, if you boot up more than one EC2 instance, you had no control where it resided--it could hypothetically be sitting on the same machine because there is no notion of location or proximity," said Adam Selipsky, vice president of product management and developer relations at Amazon Web Services.

"Now we're exposing that as a feature and you can choose to instantiate your 'nth' server in a different availability zone," he said.

Amazon Web Services last month suffered a multi-hour outage to its Simple Storage Service (S3), which affected several Web 2.0 sites.

Selipsky said the new feature will let developers add redundancy in the "vast majority" of cases.

Amazon currently gives developers the option of deploying their S3 data either in Europe or the United States.

Selipsky said Amazon will add more "granularity" on the choice of location for data over time.

Also on Thursday, Amazon Web Services introduced a IP service, called Elastic IP, that lets developers have an IP address associated to an account, rather than a physical machine.

The change makes EC2 better suited for Web application hosting, Selipsky said.

The growing buzz around cloud computing sounds eerily familiar to the utility computing noise of a few years ago.

But there is a difference--or at least people in the business swear there is.

In reading blogs and speaking to a few people, I've found that cloud computing does indeed rehash some of the ideas of the past, but there are significant technology advances, notably virtualization, that set it apart.

Cloud computing: looking for that data center in the sky

(Credit: IBM)

One such cloud company, Elastra, launched on Tuesday its server for public and private clouds. Executives at another start-up, 3Tera, told me on Monday that later this year their AppLogic "grid operating system" will let people use a single tool to deploy applications that run on a mix of Linux, Windows, and Solaris machines.

What's the link between cloud and utility computing? For starters, there's pricing. With utility computing, you get computing "power" when you need it and you pay based on consumption.

This makes sense for people who don't want to invest in building their own data centers or don't want to buy all the hardware required to handle "peaks" in demand. You can rely on outside provider to give you servers, storage, and bandwidth as you need it.

Cloud computing builds on that pay-per-drink model but makes it easier to deploy and manage applications at remote hosting providers, according to adherents.

Perhaps the most high-profile company now doing this is Amazon Web Services, which is aimed primarily at tech start-ups that like the low up-front cost.

(Credit: Forrester Research)

As Microsoft builds out its own "cloud infrastructure," it's very likely the company will offer similar hosted computing services, such as hosted storage and servers. Last month, it announced a hosted database offering. Meanwhile IBM, Sun Microsystems, and Hewlett-Packard all offer some variation of hosted utility computing.

To run an outsourced computing service effectively, service providers need automation tools. After all, there's a limit to how many servers a single person or small group of people can manage; productivity tools for IT administrators help operations scale.

A whole wave of companies were formed in the beginning of this decade to automate various tasks of data center operations, from provisioning servers to managing configurations.

Memorable "exits" from this wave of start-ups were HP's purchase of Opsware for $1.6 billion and BMC's acquisition of BladeLogic earlier this month for $800 million.

These go to '11'
But now there are new start-ups promising to do that data center automation better than before and make it easier to deploy applications out onto the Internet "cloud."

In a recent study, Forrester Research identified 11 companies that it called "cloud providers." The list includes hosting companies like Rackspace, as well as development platform providers like Salesforce.com. Systems management expert John Willis has also created a list of cloud providers and tries to demystify the concept here.

Prominent in Forrester's list is 3Tera, a company that released its first product last year. Its AppLogic creates "virtual appliances" that define which resources an application needs in order to run. A Web server, for example, may need a certain amount of memory and storage.

This allows system administrators or application architects, using a Web-based graphical tool, to set up an application choosing from a pool of servers or storage devices made available by a hoster.

"We do believe there is a difference between utility computing and cloud computing, but we believe that they aren't mutually exclusive," said Barry Lynn, 3Tera's CEO and the former CIO at Wells Fargo. "We don't understand how you can have an effective cloud unless it's deployed on a cloud platform."

The company's own software is, of course, hosted. Lynn said he believes that the next Google won't run a data center--it will get computing from the cloud.

What's different from 3Tera's approach and previous grid and utility computing automation tools is that people can work with an entire application rather than taking a piecemeal approach, company executives say.

By contrast, previous generations of tools were built to handle specific tasks, like provisioning servers, which are needed to get an application online, said 3Tera co-founder and vice president of sales and marketing Bert Armijo. The result is that application architects can get applications deployed quickly and system administrators can handle more equipment, he said.

3Tera's AppLogic tool for deploying applications to a hosting provider

(Credit: 3Tera)

Elastra makes the same claims of making it easier to deploy databases to a cloud and manage them. EnterpriseDB is using the Elastra software to have its database hosted on Amazon.

The Elastra Cloud Server creates files that define what's needed to run and manage an application. Dana Gardner at InterArbor Solutions has more here.

Elastra CEO Kirill Shaynkman, who founded portal company Plumtree, said that the software allows people to deploy databases quickly at both remote locations (hosters) or in corporate data centers.

"There are other clouds that are coming; Amazon is the most public and friendliest right now," he said. "There are private clouds already. They are just unmanageable right now."

Of course, issues associated with outsourcing remain despite all the technological progress.

When Amazon Web services went down for several hours in February, it was a rude awakening for many Web 2.0 companies that use the service without an offline backup.

But even with glitches along the way--and a healthy dose of hype from vendors--cloud computing makes it look like the idea of having computing flow like electricity is inching closer to reality.

Amazon.com Web Service's hosted storage service went down Friday morning, frustrating many Web site customers and refreshing concerns with the ballyhooed approach of cloud computing.

An online forum spiked with customer complaints Friday morning as some people found that content stored on Amazon's Simple Storage Service (S3) was unavailable or performed slowly.

The service was restored a few hours later, according to an Amazon technician. The first forum posting was timed at 5 a.m. PT, and the service was back up at just past 9 a.m.

Glitch hits Amazon Simple Storage Service.

(Credit: Amazon)

The glitch sent a ripple through the blogosphere as Web entrepreneurs, who are increasingly using Amazon's hosted computing services, pondered whether they needed a back-up plan or a more traditional hosting provider.

On the forum, some people complained about how the service glitch essentially put them out of business temporarily.

"My new sites hosts over 25,000 images on Amazon and I wake up to notice major issues this morning," wrote one customer. "The S3 service is great but this just proves you can't rely on it, this is a major issue especially since it's been down for so long."

That particular Web site operator managed to put an in-house back-up system online, a job which should be automated, according to the forum posting.

This isn't the first time Amazon has run into problems keeping its computing services running without fail.

But since launching the services, Amazon has aggressively wooed Web 2.0 start-ups as customers, many of which operate on the back of Amazon's computing infrastructure.

Photo-sharing site SmugMug, for example, said it has saved hundreds of thousands of dollars by unplugging its servers and storage gear and using Amazon's per-usage services.

Amazon said it will try to provide more technical detail on the problem, once it has been fully resolved.

John Anderson, co-founder of greeting card site SquidNote, said his customers were greeted with empty cards this morning and immediately started contacting the company. He said SquidNote's brand took a hit from the outage.

"Perhaps if AWS (Amazon Web Services) gets huge, these outages will more transparently reflect on their brand too. It might get easier for us to say, 'Hey, it's Amazon's fault...just go look at Twitter and all these other sites...see, they're down too,'" he said. "Until then, however, we just need to accept that using AWS involves a real risk to our own brand."

In another case, voting application provider PollDaddy was knocked offline for two or three hours today, which had a material impact on the company. (Disclosure: Webware will be using the Web site for its Webware 100 voting).

PollDaddy CEO David Lenehan was disappointed because S3 was pitched as bullet-proof. When the company's own servers had problems before they went to S3, the site was typically only down for a short time.

The bill from S3 won't be that much money--$4,000-$5,000--but Lenehan's confidence has been shaken. "We could get a couple of high-end servers ourselves for that," he said.

--with additional reporting from Webware.com's Rafe Needleman.

Just about everyone else on the Internet has written on the potential acquisition of Yahoo by Microsoft for $44.6 billion, but I thought that I would weigh in on what I think this might mean for search and Web services.

According to ComScore's search share numbers for December 2007, Google has 58.4 percent of the market share, with Yahoo and Microsoft trailing at 22.9 percent and 9.8 percent, respectively. If Microsoft and Yahoo combine forces and change nothing, that will put them at 32.7 percent to Google's 58.4 percent. While those numbers are certainly not enough to overthrow Google, maybe the combined minds at the two tech giants can come up with something. Somebody has to try to make a stand, so that Google doesn't run away with the industry completely. That said, I think that Google is here to stay, even though this may be its biggest challenge yet.

On the Web services side of the issue, this acquisition is looking really good for Microsoft and Yahoo. Long Zheng has a great rundown of the services that Microsoft and Yahoo provide and where they overlap. The combined user bases of Microsoft and Yahoo's Web mail services far outpace that of Gmail (they actually both beat Gmail individually), so we will put one in the win column there. If Google Talk wasn't dead enough before, it sure will be now. Google has not even come close to touching either Microsoft or Yahoo in the instant-messaging market. One more thing on IM, if this acquisition goes through, a little service called AIM is going to finally be in their sights.

Microsoft will benefit from taking control of the leading photo-sharing site, Flickr, since its only photo-sharing solution that currently exists is through its Windows Live Spaces product. Several services from the two companies will likely be merged down the road, such as Upcoming integration in Live Events, Yahoo Widgets being integrated into the Vista Sidebar, and a merger of Yahoo Answers and Live QnA. Services that are likely to get the axe include Yahoo Maps, since Microsoft's Virtual Earth technology far outperforms Yahoo's, Yahoo 360 (or whatever it ends up being once it is done "transitioning"), and most likely some of Yahoo's music services.

$44.6 billion is a lot of money to pay for an acquisition, but Microsoft has deep pockets. The deal initially makes me a little nervous, but after thinking about it for a little while, I am feeling better about it. I'm not convinced that Microsoft will take away Google's search crown as a result of this acquisition, but it will instantly become a larger player. This acquisition will lead to almost complete dominance by Microsoft over Google in the Web services arena, but not necessarily in search.

EnterpriseDB has ported its Oracle-compatible PostgreSQL database to a new platform: Amazon.com's hosted compute cloud.

The company on Tuesday started taking invitations for a beta program for EnterpriseDB Cloud Edition that will launch in March. The final product should be available this summer, according to EnterpriseDB Chief Technology Officer Bob Zurek, who spearheaded the initiative.

Amazon already offers a hosted database, called SimpleDB, but Zurek said that its database is designed for transactions and industrial-strength applications.

The service works with clustering software from Elastra, which means that servers and storage are quickly brought online to meet changes in computing demand, he explained. It taps into Amazon's Web services for hosted servers and storage, called Elastic Compute Cloud (Amazon EC2) and Amazon Simple Storage Service (Amazon S3).

EnterpriseDB has not announced pricing, but the company is looking at usage models in which customers pay for use of the database by the hour or by the month.

Potential customers are either enterprises that have custom applications that run on EntrerpriseDB or Web 2.0 start-ups that only want to pay for computing power as needed, Zurek said. He said tests show that application performance over the Internet does not degrade substantially.

He argued that the offering will make EnterpriseDB more competitive against MySQL, the leading open-source database provider which is widely used by Web companies. MySQL was acquired by Sun Microsystems for $1 billion earlier this month.