Amazon Web Services has been the frontrunner in the cloud computing race for years, and it doesn't seem to be slowing down. AWS continues to make moves to keep its customers happy and entice enterprise IT, and it has an array of cloud services that gives customers loads of options to best fit their IT environments. But just because it's the incumbent cloud leader doesn't mean all IT admins are up to speed on the ins and outs of Amazon Web Services cloud. These frequently asked questions and answers can help fill in the blanks regarding Amazon and where it stands in the cloud.
Amazon Web Services continues to drink other IT titans' milkshake in the cloud.
Are businesses really putting their cloud computing faith in a Web retailer like Amazon?
In a word: Yes. Amazon Web Services continues to drink other IT titans' milkshake in the cloud. Other vendors, such as Oracle and Microsoft, continue to work to compete and fight their way to the top. AWS made number one on the Top 10 cloud computing providers list at SearchCloudComputing.com for 2010, 2011 and 2012. And despite recent, well-publicized outages, it shows no signs of losing its throne as cloud computing king.
What types of computing resources run on Amazon Elastic Compute Cloud (EC2) and how can a CTO determine the best fit financially?
Three types of computing instances run on top of Amazon EC2, each with its own pricing structure: on-demand instances, reserved instances and spot instances. To find out the value for your company, first determine which of these three instances is the best fit for your IT environment. With on-demand, as the name implies, you only pay for the time you use the server; there is no commitment. CTOs can purchase reserved instances ahead of time for a discounted hourly rate, generally when undergoing a larger job. Spot instances were created as an incentive for companies running jobs during times of underutilization. Spot instance pricing fluctuates according to demand and compute capacity.
There has been a lot of talk about price cuts among cloud providers, including Amazon Web Services, so it's important to research what your computing needs to find the best financial fit with Amazon EC2 instances.
If I over-purchase compute capacity through AWS, will I have to consume those losses?
Purchasing on-demand, reserved or spot instances can mean you're left with computing capacity for which you have already paid but will not use. And this can only make the decision to go with on-demand instances or paying upfront for instances at a discounted rate a difficult one. However, the Amazon Reserved Instance Marketplace, which is designed to help customers recoup the losses from over-purchasing instances, makes cloud capacity planning flexible and a bit less worrisome.
What are good candidates for Amazon EC2 spot instances?
When choosing among Amazon EC2 instance types, you have to take into account performance expectations. Time-sensitive applications with little end-user interaction generally are good candidates for spot instances. Such applications could include big data analytics and testing. However, you can encounter some problems running big data analytics on AWS spot instances. If the data you intend to analyze doesn't already reside on public cloud, the costs of migrating could negate any cost benefits from using spot instances.
Incidentally, AWS offers its own cloud-based Hadoop service for big data analytics, called Amazon Elastic Map Reduce.
What application-programming interface (API) does Amazon Web Services cloud support?
AWS has its own API, and that API is the de facto standard in the cloud market. However, the battle between the AWS API and other cloud providers' API is raising enterprises' concerns about vendor lock-in and portability. Citrix and Eucalyptus, among others, use the Amazon API, but other cloud providers are teaming with OpenStack on the other side of the API war.
Caitlin White is associate site editor for SearchCloudComputing.com. Contact her at email@example.com.
This was first published in October 2012