As the pace of data generation increases, the amount of that data we want to store for some period will also increase. At some point (if it hasn’t happened already), the volume of data we want to store will exceed our manufacturing capability for hard drive (mechanical and solid-state). Now is probably the time to be thinking about how we re-introduce tape libraries into our data processing stacks.
Given how much faster the data streaming rate is for data that comes off of tape versus data being pulled from a hard drive, I wonder how difficult it would be to create a MapReduce job which gets its input from locally attached tape drives instead of traditional storage. If up front thought and consideration for such processing were done, I think it would be a very interesting experiment.
Tape is the oldest computer storage medium still in use. It was first put to work on a UNIVAC computer in 1951. But although tape sales have been falling since 2008 and dropped by 14% in 2012, according to the Santa Clara Consulting Group, tape’s decline has now gone into reverse: sales grew by 1% in the last quarter of 2012 and a 3% rise is expected this year.