Amazon Stretches Elastic Computer Cloud
Page 1 of 1
Amazon.com (NASDAQ: AMZN) has expanded the features offered in its Elastic Compute Cloud (EC2), a Web service that provides virtual storage farms for developers.
The new features are Elastic IP addresses, and the ability to launch compute instances in multiple availability zones -- what amounts to further virtualization of the already virtualized storage service.
The new capabilities make the EC2 service more robust and cut the amount of time developers spend waiting for processes to be worked through.
The Elastic IP
Elastic IP addresses are static IP addresses that can be dynamically remapped on the fly to point to any compute instance in a developer's Amazon EC2 account. This lets developers use their time better.
"If you fire up an instance of EC2 from outside and tell your DNS what your IP address is, then that instance goes down, you have to bring up a new instance and a new IP address, update your DNS and then you have to wait for that to propagate around the world," Mashery.com CEO Oren Michaels told InternetNews.com.
IP addresses are critical for Mashery.com and, with elastic IP addresses, "we never have to change our DNS so we can offer a more reliable IP address," Michaels added. Mashery.com offers an API management service to clients worldwide.
"This is a powerful feature, and they took the time to get it right," said Bob Zurek, chief technology officer of EnterpriseDB.
Zurek describes EnterpriseDB as "the Red Hat for PostgreSQL. "You can change things around and remap servers and it doesn't matter." PostgreSQL is an open source database.
Multiple availability zones
This feature virtualizes the EC2 virtual servers and gives small developers the same ability large companies have -- to distribute an application across multiple locations programmatically.
Developers only have to change a parameter in an API call.
Each availability zone runs on its own physically distinct, independent infrastructure and has independent networking, power and cooling.
This provides business continuity and ensures high availability. It also lets developers build fault-resilient Web applications by running them in multiple availability zones.
"Essentially they've now created multiple physical centers for us to use, in Virginia," Oren said, adding that the zones are "physically separate in terms of networking as well as fire and flood prevention."
While the physical servers are geographically dispersed in "a fairly small area," in the future there will be multiple geographical zones as well, so "you can even pick particular instances that are geographically different," Oren said.