Another thing to consider when partitioning the server by processes is that the memory footprint of those interpreters are non-trivial, and they'll have to be duplicated among all the processes.
1_1
Until now, we've just talked about serving up static pages. Serving requests doesn't only just require loading data from the disk, it also involves performing computation. There is quite a large amount of logic and queries that need to be executed when serving up a web page.
anonymous
This slide reminds me of assignment 4. For a web server, some requests are computing intensive (dynamic web content), others are memory intensive (static web content). In A4, we assign these two kinds of requests to the same worker node since they utilize different kinds of resources in the node(IO bandwidth/memory and CPU).
Another thing to consider when partitioning the server by processes is that the memory footprint of those interpreters are non-trivial, and they'll have to be duplicated among all the processes.
Until now, we've just talked about serving up static pages. Serving requests doesn't only just require loading data from the disk, it also involves performing computation. There is quite a large amount of logic and queries that need to be executed when serving up a web page.
This slide reminds me of assignment 4. For a web server, some requests are computing intensive (dynamic web content), others are memory intensive (static web content). In A4, we assign these two kinds of requests to the same worker node since they utilize different kinds of resources in the node(IO bandwidth/memory and CPU).