March 26, 2019 –
Title: Memory Access Efficiency in Distributed Atomic Object Implementations
PhD Candidate: Theophanis Hadjistasi
Major Advisor: Dr. Alexander A. Schwarzmann
Associate Advisors: Dr. Alexander Russell, Dr. Laurent Michel, Dr. Chryssis Georgiou
Day/Time: Tuesday, March 26th, 2019 1:00 p.m.
Location: HBL Class of 1947 Conference Room
Distributed data services use redundancy to ensure data availability and survivability. Replication masks failures, however it introduces the problem of consistency because operations may access different object replicas, possibly containing obsolete values.
Atomicity is a venerable notion of consistency, introduced in 1979 by Lamport. To this day, atomicity remains the most natural type of consistency because it provides an illusion of equivalence with the serial object type that software designers expect. Due to the strong guarantees, atomicity is more expensive to provide than weaker consistency guarantees.
Consistent storage systems continues to be an area of active research and advanced development, and there are good reasons to believe that as high performance memory systems with superior fault-tolerance become available, they will play a significant role in the construction of sophisticated distributed applications. The demand for implementations providing atomic read/write memory will ultimately be driven by the needs of distributed applications that require provable consistency and performance guarantees.
We deal with the storage of atomic shared readable and writable data in distributed systems that are subject to perturbations in the underlying distributed platforms composed of computers and networks that interconnect them. The perturbations may include permanent failures (or crashes) of individual computers, transient failures, and delays in the communication medium. The ultimate goal is to devise new algorithmic approaches that lead to efficient survivable distributed storage implementations with provable performance and correctness guarantees.