Suspend and resume an Amazon virtual computer?

Virtual servers, where you can think of a cloud of services where you pick and choose what you want and expand and contract your capacity in minutes, are becoming - I'd say - the premier choice for many outfits. There are a number of well-reputed providers, such as Amazon and Linode, but in the case of Amazon the smallest server is still rather big and pricey for being used as a alternative desktop or for throw-away development servers.

However Amazon's storage prices are low enough that having a number of standby servers, ready to recreate their state in seconds, could lend it to new applications, such as desktop machines and a company having a slew of development, staging and testing servers.

With the advent of persistent storage of your operating system on Amazon EC2 (September 2009) and now the brand new possibilty to use your own kernels, it ought to be possible for a virtual computer on the Amazon grid to be shut down with its state intact, saved to disk (i.e. hibernate), and the be woken up at later time, and you can just continue working where you are. There are such facilities in the Linux kernel, and there is also the Tux on ice project. Tux on ice allows you to save the RAM state to a file on the file system, so no need for a persistent swap partition.

This Feature Guide below is designed to teach System Administrators and other IT professionals how to utilize the User Provided Kernels in Amazon EC2. With this feature, Amazon allows you to load a para‐virtual Linux kernel within an Amazon Machine Image (AMI) or Amazon EBS volume. You can also now seamlessly upgrade the kernel on Amazon EBS‐backed instances.

Läs mer: Amazon Web Services Developer Community : Enabling User Provided Kernels in Amazon EC2