SHADE is a suit of algorithms based
on the DHS infrastructure. It exploits DHS in
order to (i) build several important histogram
structures which have proved crucial for data
management in centralized data servers and (ii)
develop a number of algorithms for computing several
aggregation queries -- for example, count, count-distinct,
sum, average, etc.
With SHADE it becomes feasible to execute efficiently
aggregation queries over highly distributed data
and to build histograms for such data, which in
turn facilitates efficient data management (eg
query optimization techniques developed for centralized
data bases, utilizing specific histogram types)
for such distributed data.
The source code of the implementation of SHADE over
FreePastry and FreeDHS, coined FreeSHADE, can be found here.
|
