Storage Technology - Scientific Data Management - page 4

1.7 Emerging Technologies ................................................ 25

1.7.1 Massive Array of Idle Disks (MAID) .......................... 26

1.7.2 FLASH ......................................................... 26

1.7.3 MRAM ......................................................... 28

1.7.4 Phase-Change Technology ..................................... 28

...........................................

1.7.5 Holographic Storage 28

................................

1.7.6 Direct Molecular Manipulation 30

1.8 Summary and Conclusions ............................................ 30

Acknowledgments ........................................................... 33

References .................................................................. 33

1.1 Introduction
Computational science is at the dawn of petascale computing capability, with
the potential to achieve simulation scale and numerical fidelity at hitherto
unattainable levels. However, harnessing such extreme computing power will
require an unprecedented degree of parallelism both within the scientific ap-
plications and at all levels of the underlying architectural platforms. Power
dissipation concerns are also driving High-Performance Computing (HPC)
system architectures from the historical trend of geometrically increasing
clock rates toward geometrically increasing core counts (multicore), 1 lead-
ing to daunting levels of concurrency for future petascale systems. Employing
an even larger number of simpler processor cores, operating at a lower clock
frequency, is increasingly common as we march toward petaflop-class HPC
platforms, but it puts extraordinary stress on the Input/Output (I/O) sub-
system implementation.
I/O systems for scientific computing have unique demands that are not
seen in other computing environments. These include the need for very high
bandwidth and large capacity, which requires thousands of devices working
in parallel, and the commensurate fault resilience required to operate a stor-
age system that contains so many components. Chapter 2 will examine the
software technology required to aggregate these storage building blocks into
reliable, high-performance, large-scale file systems. This chapter will focus
on the characteristics of the fundamental building blocks used to construct
parallel file systems to meet the needs of scientific applications at scale.
Understanding the unprecedented requirements of these new computing
paradigms, in the context of high-end HPC I/O subsystems, is a key step
toward making effective petascale computing a reality. The main contribution
of this chapter is to quantify these tradeoffs in cost, performance, reliabil-
ity, power, and density of storage systems by examining the characteristics
of the underlying building blocks for high-end I/O technology. This chapter
projects the evolution of the technology as constrained by historical trends to