Information timeliness is a major consideration for weather
tracking systems. Weather changes every
second of every minute of every day and these changes are crucial, especially
when severe weather is involved. There
are many aspects of weather including temperature, precipitation, barometric
pressure, and wind gusts and directions.
All of these dynamic aspects are important to the overall description of
weather and in providing analytical data.
Weather is different across all geographical locations but by observing
weather, patterns can be uncovered to aid in predicting weather in other
geographic locations or in the future.
Because of all of these factors, weather data is very important in many
aspects and is considered very time sensitive data.
Due to the fact that the analytical data provided by weather
is ever changing, the timeliness in recording it is very important. Applications which gather the weather data
must be very powerful and work nonstop to continue recording new data. A system of this dimension must also be very
robust in order log the many types of data from so many different regions continuously. The data warehouse which stores this
information should be very robust as data needs to be updated into the database
continuously. Because data from weather
is used to predict weather patterns and universal weather trends, it also
becomes very important that data is never lost or destroyed. A data warehouse for weather data needs to be
redundant and highly available.
Disaster recovery and backups also become very important in analyzing
a data warehouse for weather data. It
has been stated that it is crucial for the data to be constantly updated to
ensure that all data is recorded but their must also be a way to ensure that the
data is never lost or destroyed. This is
done through developing an effective back up plan and high availability
disaster recover solution so that data could quickly be recovered in the event
of a disaster or failed hardware. To do
this hardware should be configured to perform frequent backups throughout the
day to a backup data warehouse onsite which could act as a failover in the
event there would be a hardware or software failure on the production system. The next aspect that should be covered is
that data would then be backed up off site to a co-location where in the event
the primary location was destroyed the data could still be recovered. This frequency of updates and backups would
allow all of the crucial weather data to be constantly available and never lost.