Difference clocks: a new scheme for logical time in distributed systems

Logical clocks and vector clocks were proposed in the past to capture causality between events of different processes of a distributed computation. However, these clocks could not capture intraprocess concurrency. Later, bit-matrix clocks and hierarchical clocks were developed to capture interprocess concurrency as well as intraprocess concurrency. The major disadvantages of these clocks are the associated storage and communication overheads. To overcome these disadvantages, the authors introduce the concept of difference clocks. The difference clocks maintain minimal information about the differences among various local clocks. It is shown that they can be used to reconstruct the bit-matrix clocks and result in substantial reduction in storage space and communication overhead