Thursday, October 10, 2013

CLOUD COMPUTING AND DISASTER INFORMATION MANAGEMENT IN INDONESIA

Dr. Andi E. Sakya, 
Director General of Agency for Meteorology, Climatology and Geophysics, Indonesia, writes about data integration challenges faced by the agency and how cloud computing could be a possible solution.

The Agency for Meteorology, Climatology and Geophysics (in bahasa Indonesia, BMKG) collects data on weather, climate and earthquakes from more than 200 observation stations all over Indonesia. Weather, climate and earthquake data is dependent on geographic position, is time sensitive, and often relates to global phenomena. Effective collection of this data enables BMKG to provide weather, climate, earthquake and tsunami information services to Indonesian government departments, businesses and citizens.
In addition to these above main functions, recently BMKG has been appointed by UNESCO as one of the Regional Tsunami Service Providers together with India and Australia, and also appointed by World Meteorological Organisation as a Tropical Cyclone Warning Centre (TCWC).
Integrating multiple datasets
To give an example of the volume of data that BMKG processes - every two hours each of the 200 observation stations transmits raw data through telecommunication system to the headquarters in Jakarta. At the headquarters the data was stored in the database, and then integrated and processed with the data from the other observational stations and combined with data from other observation equipment such as satellite, radar, and automatic weather stations to produce weather and climate predictions.

Both daily operational as well as international assignments require BMKG to work 24 hours a day, seven days a week. As required by law, the information we create has to be disseminated quickly, accurately, broadly, and comprehensibly.
The 2004 tsunami highlighted both the importance of the information created by BMKG, and also the necessity of getting this information into the hands of policy makers quickly. Back then it took more than an hour to share earthquake data, but since the Indonesian Tsunami Early Warning System (InaTEWS) was set up in 2008 it now takes less than five minutes. By 2015, further improvements to data transmission, storage, and integration are anticipated with the Integrated Support for Early Warning Systems.
Addressing challenges with cloud computing
Our challenge in establishing such an integrated data centre lies in the fact that the time periods of measurement for weather, climate, earthquakes and tsunamis are quite different - a tsunami’s period of occurrence is around 200 years, whereas seasonal climate is about 30 years and weather is measured every two hours. Therefore, we need to be able to handle historical datasets and new data simultaneously.

Moreover, being located on the Ring of Fire, Indonesia is prone to multiple forms of natural disasters, with tsunamis occurring once every two years. Currently, we have two data recovery centres in Denpasar, Bali, one each for the TCWC and InaTEWS.
Cloud computing technology seems to pose an interesting solution to the problems above. It provides an elastic and affordable repository for data integration, while its distributive nature overcomes Indonesia’s limitations of geography.
From the operational point of view, cloud computing allows multiple users to easily feed in data in real-time. Furthermore, interoperability of cloud computing facilitates the various types of existing systems to operate together.
On the other end, it also widens possibilities and provides mobility to our information delivery channels. The information we collect is made available to the public directly through our website and through social media. Indonesia being an archipelago of 17,000 islands, it is crucial that citizens can access the information from any location, at any time and from any kind of device that they use.
Since our agency operates 24/7, implementation of the new cloud-based system without distracting our daily operation is important, and a major challenge that remains to be solved.


0 comments: