The LHC or the Large Hadron Collider is a circular particle accelerator that was designed by CERN. The collider is 17 miles in circumference and is located 574 feet below the French-Swiss border. It is also the world’s most complex and the most expensive experiment. The LHC is designed to collide lead nuclei at speeds close to that of light. Two beams of particles travel in the opposite direction and collide at 4 points where the detectors are placed.
These collisions create conditions similar to those that existed moment after the Big Bang. Under such conditions, scientists believe that they can confirm or dismiss the existence of the Higgs Boson. These collisions generate vast amount to data and to process and analyze it, CERN has created a computing grid called the Worldwide LHC Computing Grid.
The gird is composed of public internet notes and an array of private fiber optic links. The grid is distributed across 34 countries and has around 200,000 processors and a storage capacity of 150 petabytes (150 million gigabytes). The grid has three tires – Tier 0, Tier 1 and Tier 2.
The Tier 0 network is CERN’s counting room that handles the raw data from the detectors. The detectors generate around 300 gigabytes of data every second and the Tier 0 network’s CPU farm filters the raw data and generates event summaries. The Tire 0 network sends 10 terabytes of event summary data and 27 terabytes of raw data to the Tire 1 network at a speed of 10 gigabytes per second through the LHC Optical Private Network.
The Tire 1 network consists of 11 academic intuitions in North America, Asia and Europe. The Tire 1 network is as a data repository and also reprocesses the data when the LHC is calibrated. The Tire 2 network consists of around 150 institutes and is connected to the Tier 1 network through the public internet. LHC also uses distributed computing through LHC@home to process the data.
The way CERN manages the data is pretty staggering and this sure paves the way for new data management techniques.