Science Requirements and Implementation - South Pole Station Computing and Communications

A Report to the National Science Foundation
by
The South Pole Science Users' Ad Hoc Committee on Computing and Communications
William Smythe, Larry Lutz, and Robert Loewenstein

The Amundsen-Scott South Pole Station is rapidly becoming a major scientific facility. However, unlike many facilities that support a single scientific discipline, the station resources must accomodate a variety of disciplines from seismology to astrophysics. This report attempts to address the growing scientific demands on the computing and communications infrastructure. Almost all science experiments will benefit from a more effective (easier to use) link to CONUS, increased bandwidth, and support for familiar software environments. The report consists of two sections:

-A general set of guidelines for upgrading the South Pole computing and networking
environment

- Support for 4 major computing environments (SUN, VAX, Macintosh, PC)

- Non-centralized computing model

- Local Area Networks

-A specific proposal for

- Increasing the total bandwidth from South Pole to CONUS

- Implementing an Internet link to the station for the 1992-93 Austral summer season

Part I. Computing

The computing systems at South Pole must address a variety of requirements from heterogeneous science groups having great diversity in home institutions, scientific needs, and university computing environments. In addition, the South Pole systems must provide access to computing capabilities which are current enough to be useful and mature enough to be operable within the constraints of long lead procurement and shipping times, limited access by factory personnel, and limited repair facilities and spares.

There are two basic categories of science users - those at Amundsen-Scott for a few weeks, and those present for a year. The short-term users have limited time and resources with which to adapt to differences between facilities available at South Pole and those at their home institution. The long term users have sufficient time to more fully utilize specialized resources at South Pole. The largest numbers of users, by far, are the short- term summer scientists. In addition, both of these categories can have both expert and novice users, having rather different support requirements.

The near-term (5 year) requirements must be tempered with both prediction of the evolution in computing and recognition that functional research systems require some stability in the environment. The assumptions which are implicit in the requirements which follow include:
1