Systems research, more often than not, is an art of striking the right trade-off between different system design objectives. Non-trivial challenges arising from---complex dynamics of evolving computing use cases, time (era) dependent costs of computing resources, and a simplified layered model representation of complex modern day computers---are bread and butter for a system researcher. In this talk, we will revisit memory deduplication, a well known technique to improve memory efficiency in virtualized systems, from system research perspective .This talk will provide a brief historical account of memory deduplication, from its glorious adaptation in most virtual machine monitors (during 2002/03) to current times where it is in the brink of its sad demise! We will try to explore answers to the following questions: What has changed in last 15 years? What are some important challenges in memory deduplication today? Is there any scope of addressing the challenges by applying techniques across the CAOS spectrum?