Organizing your core servers is an art left for your server administrators or hosting providers. In many cases, this type of resource is being centralized and done on a massive scale "in the cloud". In the cases where you host this content yourself, plan for dual sourcing for your data in order to provide load balancing and shared resources where possible. If one channel goes down, have the alternative pick up the load and vice versa for redundancy. Systems do fail so you will thank yourself if you keep this in mind.
Organize your file directories on your web servers so that they are well indexed and stored in local server memory where possible for speed of access. This is also true for you database tables. Be sure to structure your data so your indexes are limited in size and appropriate hashing tables are used to access your data. The issue here is growth. While your applications and servers may work fine when they are first released, growth will lead to performance problems. Table indexes grow with each addition. WIth a simple linear structure, the seek time to find will grow longer and longer on average. Hashing tables can help alleviate this but changing structures in mid stream can be very painful.
Recently, database vendors have been introducing in memory improvements to their software. Just like in memory file systems, this will speed up the seek time to find your data but it will come with a price. The economics will depend on the scale of your system.
This Web site is aimed at capturing some of the basic concepts I use when creating integrated marketing programs and leveraging the internet.
Phil Gibson ©2016