Challenges for Big
Systems
As the term Big System covers a broad scope and includes many different
communities, the summit focused on a set of challenges where ontology
can provide value. Ontology can help not just model Big Systems and
their components, but also guide how they can be connected to one
another and combined to create new systems and services.
Big Data
One topic that has seen a lot of attention over the past couple of
years, especially as the cost of computing has continued its
exponential decline, is Big Data. While much of the buzz surrounding
the Big Data wave has focused on figuring out how to handle its large
volume, there are several other problems that must be addressed.
What happens when a molecular biologist wants to combine their big data
with that of a geneticist? What if we want to combine several maps, say
add Wikipedia entries to landmarks or pictures uploaded by users?
Designing Big Systems
To do any of these, we need to understand what the data is, that is, we
need semantics. To successfully combine Big Data, we need meaning and
this is where ontology comes in. Ontology provides a way of capturing
the relevant meaning of the data, transforming it into knowledge and
adding value.
Another topic that looms large for Big Systems is in their design and
construction. As we engineer Big Systems, we increasingly rely on
developing models of the system and its components. A large engineering
project, say the development of an airplane, nuclear power plant, or
car involves many people, often dispersed over large geographic areas
working together on the same system. Similarly, groups of people
striving to understand Big Systems such as our Earth, its climate, our
bodies and so on, also rely on collaboratively building and combining
models of both the entire system and its parts.
To do so effectively means that we need to build an ontology of
Systems. What exactly is a system and how are its components related to
one another and to the whole? The 2012 Ontology Summit dedicated a
track to investigating this question, with close participation of the
Systems Engineering and Modeling Languages communities.
Integrating Systems
A recurring theme in the two topics above is that of integration and
federation. In a decentralized system such as the Internet, it is
rarely possible to design from the top down. Rather, many organizations
seek to build systems, applications and services that follow a
federated model. They strive to integrate multiple, largely
independently created data or systems into a coherent whole.
To do so, it is essential that the intended meaning of a model or a
system component is communicated, otherwise costly errors may arise.
One such extreme is evidenced by the billion dollar mistake made by the
European teams making the Airbus380, where two contradictory
interpretations of “holes” caused thousands of kilometers of pipes to
be recalled and halted production midstream.
Interdisciplinary Collaboration
Lastly, a significant side-effect of all this ontologizing, is that
through explicit semantics, one of the great challenges facing humanity
today can be addressed. As the volume of our knowledge has grown so
quickly and fields have become so specialized, we’re losing a lot of
time and money to silo’d fields. Many of the problems faced by humanity
today cut across multiple fields, and very rarely does anyone one
person have the adequate skills, background or knowledge to tackle a
problem on their own.
More than ever, the ability to connect people and teams in disparate
fields, with disparate ways of looking at the world is of utmost
importance. One promising approach has been to develop ontologies to
help negotiate the different specialized languages used by different
communities. Many exciting examples abound today, from the Open
Biological and Biomedical Ontologies, to Sage Bionetworks, to the
European Union’s FutureICT project. Each of the projects aims to
provide the necessary tools and infrastructure to connect people with
different skill sets, different fields of knowledge to meet the
challenges we face today.
The 2012 Ontology Summit provides a unique forum where many of these
challenges for Big Systems were discussed and progressed. Engineers
talked about their problems and needs, while ontologists provided
suggestions for better modeling or understanding of system components.
Big Data projects shared how they were successfully using ontologies,
and where more effort was needed. Various communities presented
different approaches to tackling the challenge of integration on the
web, while others discussed how a meaningful notion of quality could be
develop.