Communication systems have been evolving very successfully. In the meanwhile,
big data has become an impressive phenomenon in a variety of disciplines.
Communication systems often carry the user data that heavily contribute to big
data (e.g. social networking data). The system’s PHY and MAC operation data
(e.g. rate adaptation records), however, are rarely collected or analyzed from
the cloud. They are quickly consumed within the system, and thus irrelevant to
big data phenomenon. The operation data, however, contain deep and broad
information about network, device, environments and human behaviors, and can be
crucial for the future system evolution. Diagnostics and optimization are two
fundamental ways of using the data, and a simple but interesting example is
considered for a DSL system.