One approach to cooling a computing system is to force air to flow over/around the components or a heat sink that is thermally coupled to the components. In many systems, this air-cooling approach is used to cool the pluggable modules.
However, in some systems air cooling of pluggable modules is suboptimal. In particular, in some systems air is not capable of sufficiently cooling the pluggable modules to the extent desired, at least not without having to compromise on some other desired parameters (e.g., size of the computing device, fan speed and power usage, etc.). This is partly due to the fact that some pluggable modules, such as electro-optical transceivers, use a lot of power and generate a lot of heat. In addition, in some systems certain pluggable modules are disposed in locations that do not receive enough air, or that receive air that is too warm because it has already cooled (i.e., extracted heat from) other components. For example, electro-optical transceiver modules are often relegated to the back of the chassis, which usually receives air that has been pre-heated by components situated upstream in the chassis, such as high-power processor and memory devices.