Of data and dust
The traditional image of an archive is one of dusty old boxes, books and papers. When your archive is digital, dust spells disaster. An innovative environmental sensor designed and built by a CERN IT specialist has become an essential element in the Laboratory’s data-preservation strategy.
The novel air particle monitoring sensor designed by CERN's Julien Leduc.
CERN’s archive holds more than 130 petabytes of data from past and present high-energy physics experiments. Some of it is 40 years old, most of it needs to be kept forever, and all of it is held on tape cartridges (over 20,000 of them).
The cartridges are held inside tape libraries with robotic arms that load them into tape drives where they can be read and written. Tape cartridges have many advantages over other data storage media, notably cost and long-term reliability, but topping the list of drawbacks is their vulnerability to contamination from airborne dust particles; a tiny piece of grit on the rollers, reels or heads can cause scratches or holes in the tape as it is being mounted or wound on the tape drive. With tape media bit sizes being smaller than bacteria or the particles emitted by a car exhaust, any damage to the tape can destroy significant amounts of data.
After a recent minor incident, where the affected tape media were scratched and punctured, finding an air particle monitoring solution became a priority. Commercially available environmental monitoring systems typically cost around 3000 EUR (plus significant maintenance costs) but they’re better suited to pharmaceutical labs or semi-clean room manufacturing units, not data centres with high air flows caused by essential air conditioning and internal cooling fans within the computers. So, Julien Leduc from CERN’s IT department designed and built a custom environmental sensor. “It’s a very simple solution to a very common problem,” explains Julien. “Based on a Raspberry Pi computer and Arduino electronics connected to several raw environmental electronic sensors, the Data Centre Environmental Sensor (DCES) sits in an empty tape drive tray inside the tape library, sampling the air 100 times every 2.5 seconds. It can measure bursts of tiny particles or the presence of larger particles, along with temperature, humidity and air flow.”
The first of these devices was installed at CERN’s Computing Centre in August 2015 and is already an essential element in the IT department’s data preservation strategy. During the DCES development process, Julien developed a portable version that can be loaned to anyone who needs to work inside the Data Centre.
The Data Centre Environmental Sensor is simple, robust, requires no ongoing maintenance and is fully integrated with the other operational monitoring systems in the Data Centre – when the sensor detects a problem, the alarm is picked up by the technical team, who can then decide how to respond. Better still, the total cost of the equipment is less than 100 EUR.
The open-source design is freely available, which is why other data centres are already installing similar systems.
