One trend from Amazon during the past year has been to increase its supercomputing and high-performance computing (HPC) options. Earlier this year the company rolled out new virtual machine instances that support cluster computing for high input/output (I/O) workloads, which are optimized for running databases and business analytics tools in Amazon's cloud. Providers such as Cycle Computing and others specialize in HPC products based on Amazon services. Just how high-performance will Amazon's cloud get and what does that mean for customers and the use cases of its product?
And another related topic to watch is how Amazon positions itself for developers. The company has already made moves in this area, with services like Elastic Beanstalk, an application development platform. AWS seems to be one of the go-to spots for hosting applications in the cloud, but it may be trying to become a place for developers to build and host applications as well, moving it more into the platform as a service (PaaS) role compared to its traditional infrastructure as a service (IaaS) model it has focused on. Just how Amazon positions itself as an application development platform, in which it would be competing with the likes of Microsoft's Azure, Google App Engine and VMware's Cloud Foundry, is another topic to watch.
What will Amazon say about the outages?
If Amazon is the 800-pound gorilla in the cloud market, then the elephant in the room is the service outages. The company has suffered three major outages in the past two years, including one this summer following major storms that ripped through the East Coast, and another after a piece of hardware caused a cascading memory failure. Amazon officials have said they don't expect to address this issue directly at the show, but there are a variety of sessions by heavy AWS users such as Netflix officials talking about how they have architected their systems for high availability and fault tolerance.