A disaggregated server is a server that breaks up components and resources into subsystems.
Disaggregated servers can be adapted to changing storage or compute loads as needed without replacing or disrupting an entire server for an extended period of time. A server could, for example, be broken into modular compute, I/O, power and storage modules that can be shared among other nearby servers. Disaggregation of software from hardware is more common. There are a number of choices of operating systems for hardware, and it’s not necessary to buy packaged software and networking equipment for the sake of adaptability.
This modularization of resources makes it possible to tear down a server when its role is complete so it can be partially reused for a customized build for one or more other purposes. Disaggregated servers also help to prevent the need for complete server replacement to upgrade just one resource group such as processors, which is a common problem in server administration.
Connectivity of components is one of the main concerns when pushing disaggregation. This concern may be addressed through photonics (electronic communication with light). Photonics offer huge amounts of bandwidth without the fickleness, air disruption and bulk of conventional electronic cables.
Intel's push into disaggregated servers includes their MXC connector and plans for up to 64 fibers at 25Gbs per fiber.