Using High Speed Networks to Enable Distributed Parallel Image Server Systems Brian L. Tierney, William E. Johnston 1 , Hanan Herzog, Gary Hoo, Guojun Jin, Jason Lee, Ling Tony Chen * , Doron Rotem * Imaging and Distributed Computing Group and * Data Management Research Group Lawrence Berkeley Laboratory, 2 Berkeley, CA 94720 Abstract We describe the design and implementation of a dis- tributed parallel storage system that uses high-speed ATM networks as a key element of the architecture. Other ele- ments include a collection of network-based disk block servers, and an associated name server that provides some file system functionality. The implementation is based on user level software that runs on UNIX workstations. Both the architecture and the implementation are intended to provide for easy and economical scalability. This approach has yielded a data source that scales economi- cally to very high speed. Target applications include on- line storage for both very large images and video sequences. This paper describes the architecture, and explores the performance issues of the current implemen- tation. 1. Correspondence should be directed to W. Johnston (wejohnston@lbl.gov), Lawrence Berkeley Laboratory, MS: 50B - 2239, Berkeley, CA, 94720. Tel: 510-486-5014, fax: 510-486-6363; or Brian Tierney (bltierney@lbl.gov), Tel: 510-486-7381. (WWW: http:// george.lbl.gov) 2. This work is jointly supported by ARPA - CSTO, and by the U. S. Dept. of Energy, Energy Research Division, Office of Scientific Comput- ing, under contract DE-AC03-76SF00098 with the University of Califor- nia. This document is LBL report LBL-35437. Reference herein to any specific commercial product, process, or service by trade name, trade- mark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement or recommendation by the United States Govern- ment or the University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or the University of California, and shall not be used for advertising or product endorsement purposes. The following terms are acknowledged as trademarks: UNIX (Novell, Inc.), Sun and SPARC- Station (Sun Microsystems, Inc.), DEC and Alpha (Digital Equipment Corp.), SGI and Indigo (Silicon Graphics, Inc.). 1.0 Introduction This distributed data system was developed in the con- text of the MAGIC 3 gigabit testbed and our DOE program in high-speed distributed imaging systems. While much work has been done on using networks to provide the interconnect for workstation-based parallel computing sys- tems (“clusters”) (see, for example [5]), not much attention has been paid to the potential of the network to provide high-speed data storage systems. In the MAGIC testbed a high-speed imaging application motivates the work of the collaborating organizations. The general goal in MAGIC is to explore the concept of using large, on-line image archives like those at the USGS’s EROS Data Center as a source of data for a terrain visualization application that ultimately might let one “walk” or “drive” through the landscape anywhere on the surface of the Earth (or elsewhere). The application com- bines terrain elevation models with high-resolution aerial or satellite images to produce a virtual reality - type inter- action with the landscape. This type of application requires data that has been processed in such a way that the surface imagery and elevation models can be combined to form a 3-dimensional image of the landscape, algorithms that can navigate thought this virtual landscape, and a way of get- ting the image data to the application based on the requi- sites of the navigator. The combination of having large data archives that might be in many different locations, the need for high-speed data delivery (300-400 Mbits/s), and a desire to allow the application to be sited anywhere on the network, leads to the general requirement for a distributed source of image data. This requirement has led us to inves- 3. MAGIC (Multidimensional Applications and Gigabit Internetwork Consortium) is a gigabit network testbed that was established in June 1992 by the U. S. Government’s Advanced Research Projects Agency (ARPA)[13]. The testbed is a collaboration between Mitre, LBL, Minne- sota Supercomputer Center, SRI, Univ. of Kansas, Lawrence, KS, USGS - EROS Data Center, Sprint, Northern Telecom, U. S. West, Southwest Bell, and Splitrock Telecom. More information about MAGIC may be found on the WWW home page at: http://www.magic.net/ ISSN 1063-9535. Copyright (c) 1994 IEEE. All rights reservered.