BEWEB keynote speaker: Prof. Torben Bach Pedersen (from the Database and Programming Technologies group at Aalborg University, Denmark).
Title: Research Challenges for Cloud Intelligence
Abstract: Cloud intelligence is a collection of technologies emerging
from the migration of business intelligence and analytics technologies
to a cloud computing environment combined with exploiting the massive
range of new intelligence opportunities opened up by cloud computing.
With cloud computing, we have to handle a paradigm shift in the
computing infrastructure. Here, existing concepts such as storage,
processes, computation, etc., are virtualized to all be parts of "the
cloud". Thus, technologies have to become device and location
independent, e.g., different parts of the data will all reside "in the
cloud," but can be dispersed, and even migrating, between many physical
locations and devices. We thus have to start thinking about data
management as a service, which in turn entails that analytics and
intelligence also become services. Another aspect of cloud intelligence
is that every service potentially should be scalable to a global level,
meaning that massively parallel computing techniques such map-reduce and
beyond will become the standard programming model. Cloud intelligence
also entails agility, the ability to assemble the necessary resources on
demand, not only in terms of computing power, but also in terms of data
sources. In a long term perspective, all the data in the world is
available "in the cloud", some in public clouds, including the web, some
only in private clouds, and some in hybrids. Thus, the current
significant difference between own and public data instead becomes a
sliding grey scale. This development poses many challenges to cloud
intelligence: with sensitive data outsourced to a cloud provider,
privacy and security becomes essential parts of any analytics solution;
with the increasing dependence on cloud computing, reliability becomes
mission-critical; and cloud intelligence also needs to become
energy-aware to reduce the massive energy consumption in data centers.
Finally, the perhaps greatest challenge comes from fully utilizing all
the new (types of) data available in cloud, for an ever increasing range
of intelligence functionality to many more and much more diverse users.