Databases Reference
In-Depth Information
FIGURE 1.1: Inside the Large Hadron Collider (LHC) tunnel. An example of
a large-scale data generation facility (Source: CERN).
n
2
can range from O (n log n) to O
or worse, depending on the type of prun-
ing applied [6]. These types of algorithms are computationally expensive and
infeasible for large data sets.
The integration of computational devices within the Internet architecture
has seen rapid expansion in commodity use items, such as phones, and com-
modity use infrastructures, such as roads. Sensory data are generated and
used remotely to interact with the environment. This connectivity between
devices and sensory-enabled objects is commonly known as the Internet-of-
Things (IoT) and proposes pervasive computability and sensor-led control
through a plethora of smart objects, around us. These smart objects are ev-
eryday physical things that have been enhanced with a small electronic device
to provide local intelligence and connectivity to the Internet [7]. This enhance-
ment bridges the gap between the physical and information domains. With
seamless connectivity between smart objects and high-performance compu-
tational systems, such as Internet servers, it may become possible for us to
create large-scale sense-compute systems that exhibit the e
ciencies of bio-
logical nervous systems.
Imagine a household equipped with fully connected smart devices with em-
bedded sensors. These sensors would detect the level of heat emitted by each
attached device and collectively determine the heat level generated. Informa-
tion on current weather conditions obtained from the Internet could then be
used to adjust the thermostat of an air-conditioner to load-balance the heat
from all of the devices with the existing room/house conditions, thus creating
an intelligent and adaptive heat control system. To be effective, this sensing
capability requires a massive amount of data to be extracted continuously in
real-time. Therefore, a mechanism to extract these data should be considered.
In this context, we can use patterns to represent a collection of sensory data
over a specific timeframe. We can implement recognition on these patterns to
detect and adjust the level of heat required.
As another example, one can imagine wirelessly connected sensors embed-
ded along a road between two cities, primarily for the purpose of tra
c man-
agement. The sensors could signal an incoming tsunami or seismic event and
provide invaluable minutes before a cataclysmic event. Such a network could
also provide real-time data for calculating routes and arrival times.
