Talks SPTDC 2019
Achour Mostefaoui
University of NantesByzantine agreement (part 1)
This talk will discuss how to solve the consensus problem in the case where some processes of a distributed message-passing system may exhibit a Byzantine behavior.
Read more
Leslie Lamport
MicrosoftThe Paxos algorithm or how to win a Turing Award (part 1)
How to think about concurrent systems mathematically is explained using the Paxos consensus algorithm as an example. First, the problem to be solved is precisely specified. Then, a "shared memory" voting algorithm is specified and shown to implement the problem specification. Finally, the Paxos algorithm is specified and shown to implement the voting algorithm. How mathematical thinking is used in industry is then briefly discussed.
Read more
Ittai Abraham
VMWare ResearchByzantine fault-tolerance, state machine replication and blockchains (part 1)
In the first part of this tutorial we will provide a new general framework that can be used to explain many consensus protocol variants, and in the second — touch upon the deep connections between Byzantine fault-tolerance and blockchains.
Read more
Dan Alistarh
IST AustriaDistributed and concurrent optimization for machine learning (part 1)
The goal of this tutorial lecture is to provide the audience with an overview of standard distribution techniques in machine learning, with an eye towards the intriguing trade-offs between synchronization and communication costs of distributed machine learning algorithms, on the one hand, and their convergence, on the other.
Read more
Trevor Brown
University of WaterlooPractical aspects of multicore programming (part 2)
A brief introduction to concurrent data structures, including what it means for concurrent data structures to be correct, and how they should be implemented to achieve high performance on real systems.
Read more
Michael Scott
University of Rochester
Eliezer Gafni
UCLAThe coordination power of distributed computing models (part 2)
Description will be announced soon
Read more
Maurice Herlihy
Brown University Computer Science DeptCross-chain deals: blockchains on ACID (part 1)
Modern distributed data management systems face a new challenge: how to allow autonomous, mutually-distrusting parties to co-operate safely and effectively. This challenge presents many questions familiar from classical distributed systems: how to combine multiple steps into a single atomic action, how to recover from failures, how to synchronize concurrent access to data, and what it means for that action to be correct. This lecture describes how each of these notions requires rethinking when participants are autonomous and potentially adversarial.
Read more
Leslie Lamport
MicrosoftThe Paxos algorithm or how to win a Turing Award (part 2)
How to think about concurrent systems mathematically is explained using the Paxos consensus algorithm as an example. First, the problem to be solved is precisely specified. Then, a "shared memory" voting algorithm is specified and shown to implement the problem specification. Finally, the Paxos algorithm is specified and shown to implement the voting algorithm. How mathematical thinking is used in industry is then briefly discussed.
Read more
Trevor Brown
University of WaterlooPractical aspects of multicore programming (part 1)
A brief introduction to concurrent data structures, including what it means for concurrent data structures to be correct, and how they should be implemented to achieve high performance on real systems.
Read more
Eliezer Gafni
UCLAThe coordination power of distributed computing models (part 1)
Description will be announced soon
Read more
Dan Alistarh
IST AustriaDistributed and concurrent optimization for machine learning (part 2)
The goal of this tutorial lecture is to provide the audience with an overview of standard distribution techniques in machine learning, with an eye towards the intriguing trade-offs between synchronization and communication costs of distributed machine learning algorithms, on the one hand, and their convergence, on the other.
Read more
Ittai Abraham
VMWare ResearchByzantine fault-tolerance, state machine replication and blockchains (part 2)
In the first part of this tutorial we will provide a new general framework that can be used to explain many consensus protocol variants, and in the second — touch upon the deep connections between Byzantine fault-tolerance and blockchains.
Read more
Danny Hendler
Ben-Gurion University of the Negev
Achour Mostefaoui
University of NantesByzantine agreement (part 2)
This talk will discuss how to solve the consensus problem in the case where some processes of a distributed message-passing system may exhibit a Byzantine behavior.
Read more
Danny Hendler
Ben-Gurion University of the Negev
Maurice Herlihy
Brown University Computer Science DeptCross-chain deals: blockchains on ACID (part 2)
This lecture describes how questions familiar from classical distributed systems require rethinking when participants are autonomous and potentially adversarial.
Read more