Amazon's Linux cloud computing out of beta, joined by Windows (cnet)

[Posted October 23, 2008 by cook]

cnet reports on the latest Elastic Compute Cloud developments from Amazon. "The Elastic Compute Cloud, a service that gives customers on-demand access to Linux servers, is now out of beta testing, said Jeff Barr, evangelist for the collection of online options collectively called Amazon Web Services. "Amazon EC2 is now in full production," Barr said in a blog post Thursday. And as promised, EC2 now offers Windows in a beta test, joining Sun Microsystems' OpenSolaris and Solaris Express Community Edition. Along with those moves, EC2 now comes with a service level agreement, a formal commitment that the service will be available at least 99.95 percent of the time."

Good for one-off testing and development

Posted Oct 24, 2008 8:00 UTC (Fri) by jhs (guest, #12429) [Link]

"Cloud computing" is definitely hyped; however, I have really enjoyed EC2 for testing software on various Linux distributions quickly and conveniently.

Time was, I had to install several distros with virtualization to have a broad development and testing environment for one-off activities. But with an EC2 account, I just choose a disk image, boot it up, and log in via SSH or X. At ten cents per hour, it's pretty cheap for to work in this way.

Most recently, I wanted to try Sun's MySQL Workbench, but its dependencies are pretty steep. Wanting neither to upgrade my laptop's distro, nor to install Cairo from source, I just booted up the latest Ubuntu in EC2 and I had the software going in no time.

The Cloud is Coming

Posted Nov 1, 2008 4:07 UTC (Sat) by dmag (guest, #17775) [Link]

Cloud Computing is very interesting. For those who don't know, Amazon's EC2 is nothing like a dedicated hosting provider. First, your data is never backed up (most people script a solution involving S3). Second, your server can fail at any time. Third, when your server fails, your data is gone. (Wiped clean. Hope you didn't need it.) They have an SLA (Service Level Agreement), but that's only about launching new instances.

This makes Amazon's job dead simple compared to an old fashioned dedicated/shared hosting provider. There are no angry phone calls when the inevitable hardware problems happen. There are no highly paid admins/developers waiting for fsck to complete. There are just guys on rollerskates to pull out the bad hardware at the end of the month.

This complicates the life of the users a bit, but there's plenty of ways to get HA (High Availability) out of multiple computers. Since the servers are unreliable, but cheap and disposable, this makes automation essential. The end result is actually more reliable (for a number of reasons I won't go into here).

Consider these trends:
- In the old days, companies had 3-5 servers. The OS and the applications were "manually" installed.
- Now, companies have dozens to hundreds of servers. OSes and Applications have to be managed at a higher level. (Try Puppet, it's neat.)

- In the old days, applications ran on 1 server. Well, 1 development server, 1 staging/QA server, and 1 production server.
- Now, people write apps that run on multiple servers. You need 5 development servers, 15 staging servers, and 40 production servers. Then you need to load test 50 servers, but just for an hour or two.

- In the old days, people bought big expensive ($30K-$50K) servers in the hope that they were more reliable. If a server went down, it was big headaches for everyone.
- Now, the smart people buy cheap ($2K) servers and tie them together with High Availability software. (Just like RAID, but applied to compute power.)

- In the old days, the OS was fussy about the hardware. (Remember driver disks?)
- Now people use Linux, or at least Virtualization.

- In the old days, highly paid sysadmins/developers were involved in setting up hardware, and diagnosing faults. They had to know which servers required EISA or ECC. They had to know which of the dozen "standard" SCSI connectors they needed. Few companies were big enough to have a dedicated hardware group.
- Now, people can pay someone else (guys on rollerblades) to manage hardware. Who cares why the server died? Does a developer care if it was a power supply vs an ethernet card? Why not just spin up a new server and get on with life?

There will be much debate about what things are appropriate to put in the cloud, but that is missing the point. Cloud Computing is here, and many people will find some use for it (once they understand it).