, , ,


My interest in today’s CITRIS lecture by Ethan Miller, in the red shirt above, entitled Pergamum: Energy Efficient, Reliable, Disk-Based Archival Storage, was in finding out if a thousand year digital data storage site using Pergamum was possible. The answer was yes, with some conditions such as a minimum boot up and restore cycle of once per year. The intention is to have as safe a storage for the data of modern humanity as it is possible to create. My goal for an Earth Ark project is to be able to recover as much of modern civilization as possible after a major atomic war. That includes: humans, animals, plants and data. All of this is ultimately information which is encoded in the form of DNA or digital data. Data before about 1995 was stored mostly on paper, but after that date most data is being increasingly kept in digital form in various ways, which unfortunately evaporates over time due to various reasons discussed last week in my blog, Ensuring Digital Documents. The storage system discussed then was based on the LOCKSS (Lots Of Copies Keep Stuff Safe) computers located at Stanford University. But all of this is lost if the materials, the machines and “disks” are lost. Thus it seems reasonable to store a couple of copies where they will survive permanently. The cold and remoteness of the southern side of the Antarctic mountains would be a great place for properly prepared materials. -80.3 -82.0

The Pergamum system appeared to be much cheaper to create for massive data storage than LOCKSS and more highly portable. It is based on a small number, about five, computer data storage systems each with complete self restoration capabilities which when they do eventually have a mechanical failure of a component can be easily reconstituted by replacing that single component, and restoring it to original condition from the other parallel sources. It is similar to RAID systems, but with the more advance self healing qualities Pergamum is much more robust. Below is a comparative slide.

Cost efficiency of Pergamon vs. Tape

Cost efficiency of Pergamon vs. Tape

To store 10PB of data would cost about $4,700 (there may have been  three more zeros in this). It’s hard for me to think in Petabytes having just migrated to Terabytes from what recently seemed like gigantic Gigabyte hard drives and within easy memory of Megabytes sized drives and not long before that Kilobytes. My friend Dell said the whole East Bay shook when he fired up his gigantic 20 Megabyte storage hard drive back in the 80s.

My goal is to find a robust system that is so independent from modern society and its current technology that it can be fired up a thousand years hence, after who knows what assaults, and without too much fuss present the data stored on it to a not particularly savvy group of human beings. This system would be stored in a cold Antarctic environment inside of a shipping iso-container and then further protected inside of another container. Ideally the system would be so simple it could be set to operate upon flipping a single ON switch or by simply opening the door. Perhaps a solar powered electric system could charge the components of the computer system and fire it up once a year when conditions were most favorable. The container might be skid dropped from a low skimming air plane or possibly trucked in to Vinson Massif. What ever works.