Amazon has given developers scope to deploy secure websites on the cheap, by increasing the number of IP addresses that can be assigned to each one of its rentable computers.
The new Virtual Private Cloud feature, announced on Friday, boosts the number of Elastic IP Addresses per rented server from two to as high as 30. This lets developers base multiple SSL-secured websites on a single instance, allows for more sophisticated network appliances and load balances, and makes moving applications easier, according to the company.
The move fits with AWS's demonstrated commitment to tailoring its products to the demands of customers and follows its decisions to introduce other sought-after technologies, like Hadoop-variant Elastic MapReduce and the Workflow service.
Developers can use one Elastic IP Address per EC2 instance (server) for free, and thereafter each additional address will cost $0.005 (£0.003) per hour. There is no charge for additional IP addresses used solely to shuffle applications within the Virtual Private Cloud. With the update, Amazon appears to have taken a lead on cloud rivals like Google and Microsoft. Documentation from both Microsoft and Google indicates that developers can only assign a single IP address per rented server. If so, this would prohibit developers from being able to mount more resilient load balancers on the services and host multiple websites on a single instance. This is probably because Windows Azure and Google Compute Engine both have platform-as-a-service (PaaS) components which, though more expensive, can handle this sort of thing for developers. The lack of multiple IPs per instance in these companies' infrastructure components may be a way of encouraging developers to fork out for higher-level platform services. As for why Amazon has announced this now, it could be due to lessons learnt from the AWS outage last weekend, which triggered software bugs that derailed the company's load-balancer technology. This hit major AWS customers such as Netflix.
Over at Business2Community, Greg Schulz's take is that the technology is being rolled out to strengthen load-balancing technology. In the case of a bad availability zone outage, this means developers can build applications that work around it.