Describes experiences and lessons-learned from the Large Synoptic Survey Telescope (LSST) project. The sheer size and complex nature of LSST, bring a unique set of challenges and a massive software modeling endeavor. UML and ICONIX process are critically important in such an undertaking. Includes an in-depth discussion of tailoring ICONIX Process to support algorithmically-intensive development.

In its first month of operation, the LSST will see more of the Universe than all previous telescopes
combined. Its rapid‐fire, 3 billion pixel digital camera will open a movie‐like window on objects that
change or move; its superb images will enable mapping the cosmos in 3D as never before.
Surveying the entire sky every few days, LSST will provide data in real time to both astronomers and
the public.

The Data Management (DM) part of the LSST software is a beast of a project. LSST will deal with
unprecedented data volumes. The telescope’s camera will produce a stream of individual images
that are each 3.2 billion pixels, with a new image coming along every couple of minutes.
In essence, the LSST sky survey will produce a 10 year “sky movie”.

For LSST Data Management, a good strategy is to make extensive use of rapid prototyping (in this
case algorithm development via prototyping) in addition to UML modeling. So a two‐pronged
strategy of prototyping and modeling has been underway on LSST for a few years now.

A further complication in modeling LSST’s DM software is that the development is being done in
multiple, independent, geographically distributed research organizations—places like Princeton,
Stanford, University of Washington and Caltech, among others. This places some significant
requirements on the modeling tools that are used by the project.