Trends
The NPDW is the basic design for the Data Warehouse Parallel Architecture Project
(DWPA, 2005), which focuses on architectural characteristics, automatic reorganization,
load balancing, response time prediction, and automatic adaptability for
the low-cost node-partitioned data warehouse. These are in line with current and
future trends on database research in related issues, which include database selftuning
and autoconfiguration (Chaudhuri & Weikum, 2002; Schiefer & Valentin,
1999; Weikum, Moenkeberg, Hasse, & Zabback, 2002). Runtime prediction is
also an important objective for current and future research on database engines.
There are very interesting recent works on runtime estimation and improvement
(Chaudhuri, Narasayya, & Ramamurthy, 2004; Luo, Naughton, Ellmann, & Watzke,
2004) that can be usefully adapted to parallel settings and in particular to the
NPDW environment. There is nowadays a market trend towards more and more
open-source software, including open-source database engines being deployed in
organizations and cost-consciousness in both hardware and software platforms is
increasingly important. In this context, the DWPA concept of an architecture that
can run anywhere efficiently and adaptively also seems to be in line with current
trends. Besides, many of the issues discussed in this chapter can also be applied to
other parallel architectures that are increasingly deployed, in particular symmetric
multiprocessors (SMP) and clusters of SMPs.
Pages:
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427