mobile2 - 12:33 am on Jun 5, 2012 (gmt 0)
Using Nginx or Appache and running normal webserver tasks (mostly php) within Amazon AWS EC2.
Now we want to add range-request streaming and there are 2 options (if there are more let me know :-)
Option 1: Have the same web servers that do the php web serving, handle mobile and developer api's also handle the range request streaming of media files. (we use c1.medium instances)
Option 2: Use separate servers for the streaming and don't combine it with webserver / api tasks.
Option 1 we think has the advantage that we manage one pool of servers.. we just add new servers and that's that...any spikes between the different tasks balance each other out.
Option 2 would require us to manage 2 pools of servers.. what if suddenly a use case evolves that spikes streaming... vs. web server tasks..we'd shut down web servers and boot up streaming servers. Option 1 we estimate those tasks would just balance each other out.
Any help/advice by someone in the know would be greatly appreciated.
Currently we run apache servers.. for range request we might switch to NginX... or keep using apache. Same question is relevant though. Not sure if the answer is different between apache and nginx.