01-31-2016, 04:10 PM
(This post was last modified: 01-31-2016, 04:15 PM by Madroxprime.)
Clustering(like a beowulf cluster) sets an array of computers to work on processes in parallel. It works by having a master node that handles handing out processes to the slave nodes over a network connection through a message passing interface, kinda using the network like a bus.
It's super useful for projects that are limited by a need for CPU cycles but not so good for games or a lot of casual usage since the network passing process messages would be a little laggy for games.
So it lets work that would take a long time, get done faster.
There's also load balancing clusters which work by kinda playing traffic cop with users trying to access data, directing them to other servers when there are too many users for any one server to handle the load comfortably.
What makes this super useful with the Pine is that the power draw is so much lower, that the financial upkeep can be significantly less.
It's super useful for projects that are limited by a need for CPU cycles but not so good for games or a lot of casual usage since the network passing process messages would be a little laggy for games.
So it lets work that would take a long time, get done faster.
There's also load balancing clusters which work by kinda playing traffic cop with users trying to access data, directing them to other servers when there are too many users for any one server to handle the load comfortably.
What makes this super useful with the Pine is that the power draw is so much lower, that the financial upkeep can be significantly less.