We have created a product that will potentially generate many requests for a data file that is located on our server. We currently have a shared hosting server running a PHP script to query the database and create a data file for each user request. This is inefficient and is not a problem yet, but we want to move to a more scalable system, so we are looking for EC2. Our main problems are the ability to handle large amounts of traffic when they occur and to provide low latency for users downloading data files.
I am not 100% sure how all this will work, but this is the idea:
We use an instance of EC2 to host our administration panel and to create files that are provided to application users. When any administrator makes changes that affect these data files (which are uploaded by users), we make a copy to S3 using CloudFront. The idea here is to get data cached and pending on S3 so that we can reduce the computation time and use CloudFront to get low latency for all users requesting files.
I am still studying the system and want to know if anyone has any feedback on this idea or how it might work. I am also interested in learning about the goals of projects like Cassandra. I understand that simply placing our application on EC2 servers makes it scalable in terms of the nature of the servers. Is Cassandra simply under-utilized, or is there a reason to use such a system, even if on EC2?
CloudFront: http://aws.amazon.com/cloudfront/
EC2: http://aws.amazon.com/cloudfront/
Kassandra: http://cassandra.apache.org/
source
share