Faced with the challenge of enormous data volumes specified by forthcoming imaging infrared interferometers, and the necessity of processing this data in a timely fashion, we have conducted design studies and built prototypes of a distributed data processing system capable of meeting throughput and latency requirements. Such a system must take in raw instrument data (interferograms of scene and reference black bodies) and produce radiometrically calibrated infrared spectra, as well as vertical atmospheric profiles retrieved from these spectra, including temperature and water vapor mixing ratio. The system must also take advantage of the inherently parallel profile of the data processing, and be designed to survive many generations of computing hardware evolution. We illustrate the current iteration of our test system, together with the problems encountered and the solutions devised. Applications to current ground and airborne interferometers as well as future spacecraft instruments are also presented.