Thursday, March 20, 2014

CloudBox can improve flight safety with cloud analytics

Over the past two weeks there has been intense search for the missing Malaysian Airlines Flight MH370 (#MH370). Every plane is equipped with the black box that often yields clue about what happened inside the flight. Usually after an accident or crash, the experts retrieve this black box and look for clues on what may have gone wrong. Ever since I learnt about this black box stuff, I had been in awe. However with this recent MH370 loss and long search for it, I wonder why are we not using real-time cloud (service) analytics to keep track of planes in flight?

There are likely several challenges with uplinking flight data and doing timely analytics. Over the past decades the cloud services industry has developed amazing computing, data, and communication infrastructure that can be used to solve problems. Let us assume that each plan can be equipped with a cloud box that can capture and feed diagnostics data and appropriate pilot + officer conversation data from the planes in flight to cloud analytics services. Let us look at the data processing capacity required. For now I will ignore the costs for cloud box and associated services. In the end, any cost can be ascribed to increased safety, reduced insurance rates, and worst case become a premium for flight charges across the board.

According to various Internet posts (one example here), there are about 100,000 flights per day.

Assume each diagnostics + conversational sample collected is compressed to 1MB per uplink.

Assume that each sample is uploaded to at least 3 satellites to ensure redundancy at 30-second intervals.

(And perhaps we may need adaptive algorithms to detect change in course, which can trigger faster uplinks at one sample every second.)


At 30-second samples, there are 2880 to be accurate, resulting in about 10 GB of compressed data per day per plane.

For 100K flights, that is about a peta-byte of data coming in per day.



 

 

Provided we have good distribution of satellites and good processing capabilities, there will be sufficient computing and bandwidth to handle such traffic. The data is very suitable for parallel processing which can be done through a variety of mechanisms. It is achievable, cost and fitting economics aside. On the ground existing and new receiver stations can process such data to provide feeds to the control towers and authorities to take action.

 Perhaps the cloud box is already in action for non-civil purposes. If not, new installations will come up in the coming years. Technology can help us tame problems.

 In the future skies, we can expect air travel to be more safer.

No comments:

Post a Comment