Seminarium: Systemy Rozproszone
16 lutego 2012, godzina 12:15,
sala
During the first part of the presentation I'll present an example distributed backup system using duplicate elimination technique for extremely good (20:1) data compression of stored information and the problem of data fragmentation caused by such solution. The second part will cover the discussions about possible solutions to the issue and presentation of the algorithm called context-based rewriting (CBR) as a way to improve the restore performance by clever modification of the original duplicate elimination algorithm with only small additional cost.
The presentation will be based on the fastest in the world backup system HydraStore and my research work on that project (within 9LivedData LLC and during my PhD studies on the University of Warsaw).
Required special abilities: