blob: f350cfd244fed0b0cc59e39ba9d98ab81b8e8c73 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
|
\section{Conclusion}
\label{sec:conclusion}
In this chapter, we discussed the creation of a dynamization system
based
upon the Bentley-Saxe method that can be used to create dynamized
sampling
data structures that outperform dynamic baselines and feature a
configurable
design space. Specifically, we discussed dynamized versions
of the alias
structure for weighted set sampling, the alias-augmented
B+tree for
weighted independent range sampling, and the ISAM tree for
independent
range sampling. In each case, the static structures were
dynamized with
support for inserts and deletes without compromising
their query performance
advantage over dynamic baselines, and while
matching or exceeding the dynamic structures' insertion performance.
The techniques proposed in chapter, however, are limited to a very
specific class
of data structures for addressing a very specific type
of search problem. While these results are promising, they fall short
of a general
solution to data structure dynamization that addresses
the
limitations of classical dynamization techniques discussed in
Chapter~\ref{chap:background}. In the next chapter, we will take several
of the results of this chapter, generalize them, and apply them to a much
wider range of data structures.
|