- Watch recording
- Watch recording
Synchronization primitives can be faster with SegmentQueueSynchronizer
This talk will present SegmentQueueSynchronizer — an abstraction for writing synchronization primitives with abortable operations simply and efficiently.
Nikita Koval
JetBrains
- Watch recording
Safe user-level sharing of memory-mapped resources
This talk will present the Hodor project, which overcomes the single address space limitation by using the memory protection keys of recent Intel processors to implement a protected library mechanism.
Michael Scott
University of Rochester
- Watch recording
Paxos vs Raft: Have we reached consensus on distributed consensus?
Which algorithm, Paxos or Raft, is the best solution to distributed consensus?
Heidi Howard
University of Cambridge
- Watch recording
CRDTs: The hard parts
In this talk Martin goes beyond the introductory material on CRDTs, and discusses some of the hard-won lessons from years of research on making CRDTs work in practice.
Martin Kleppmann
University of Cambridge
- Watch recording
Multithreading in UI application: When some threads are more equal than others
In this talk we will see specific problems and solutions related to UI applications with heterogeneous threads. It will help you to avoid many pitfalls and make familiar with several approaches and tools that are invaluable for UI applications development.
Dmitry Ivanov
Huawei
- Watch recording
Tissue vs silicone musings on the future of deep learning hardware and software
What do the properties of neural tissue have to do with the way we process data in convolutional neural networks (CNNs)? Perhaps more than we think. Nir Shavit will talk about that during this session.
Nir Shavit
MIT
- Watch recording
Scaling distributed ledgers: Achieving arbitrary throughput and sub-second latency
This talk describes how to improve on the following limitations of blockchain technologies and smart contract platforms: low throughput and high latency.
- Watch recording
Debugging data races
A (brief) primer on debugging data races — both parallel and distributed.
Cliff Click
CRATUS
- Watch recording
HTAP made live — from the engineering perspective
In this talk, Ed will talk about TiDB and its columnar storage engine TiFlash from the perspective of engineering design, and explain how TiDB utilizes and extends Raft protocol to eliminate the impact of OLAP workload on OLTP, to achieve an industrial-grade HTAP database, and give relevant benchmarks and user cases.
Ed Huang
PingCAP
- Watch recording
Programming language memory models: Problems, solutions, and directions
We consider ups and downs of modern memory models for programming languages and requirements imposed on them by languages' pragmatics.
Anton Podkopaev
MPI-SWS, JetBrains Research, NRU HSE
- Watch recording
Redundancy does not imply fault tolerance: Analysis of distributed storage reactions to single errors and corruptions
In this talk we will analyze how modern distributed storage systems behave in the presence of storage faults such as data corruption and read and write errors.
Aishwarya Ganesan
University of Wisconsin - Madison
- Watch recording
Oak: A scalable off-heap allocated key-value map
OakMap is part of the open source Oak library. It is a concurrent Key-Value Map that may keep all keys and values off-heap. OakMap decreases the memory requirements up to 3 times and experiments show that OakMap is often 2x faster than Java’s state-of-the-art concurrent skip-list. We will discuss OakMap technology and working with the off-heap memory: updating and concurrency.
Anastasia Braginsky
Yahoo! Research
- Watch recording
- Watch recording
Extending PostgreSQL to Google Spanner architecture
In this talk, we will look at the architecture of YugabyteDB that enables it to support all PostgreSQL features along with distributed transactions, resilience, scalability, and geo-distribution of data.
Karthik Ranganathan
Yugabyte
- Watch recording
Wait-free memory reclamation and data structures
In this talk, we will discuss the memory reclamation problem, existing techniques, and how a recent wait-free WFE technique is constructed. We discuss what key challenges WFE solves so that it can be universally applied regardless of the data structure organization.
Ruslan Nikolaev
Virginia Tech
- Watch recording
Designing distributed systems with TLA+
This talk will focus on one software modeling tool called TLA+, which has seen spectacular success in real-world projects. We will also demonstrate it on an example of a distributed system and closeout by discussing learning resources.
Hillel Wayne
Windy Coast Consulting
- Watch recording
- Watch recording
New locks for the old kernel
In this talk Alex will describe recent efforts to modernize Linux kernel locks, focusing on two concrete ideas.
Alex Kogan
Oracle Labs
- Watch recording
Erasure Coding at Scale
In this talk Maxim will give a brief overview of Erasure Coding schemes that we employ and discuss various real-world scenarios and lessons learned while operating these systems at scale.
Maxim Babenko
Yandex
- Watch recording
Efficient and reliable microservices
In this talk Oleg will try to cover the advantages of stateful vs stateless microservices on the example of Odnoklassniki social network, discuss how statefulness affects reliability and accessibility of services and how it helps to build faster applications.
Oleg Anastasyev
Odnoklassniki
- Watch recording
Java threads are losing weight performance review of Project Loom
Project Loom is an attempt to introduce lightweight concurrency constructs to Java. Can it change the game of Java concurrency? Let's look into details.
Sergey Kuksenko
Oracle
- Watch recording
Distributed and concurrent optimization for machine learning
The goal of this talk is to provide an overview of the role of distributed computing 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.
Dan Alistarh
IST Austria
- Watch recording
Demystifying Bitcoin
The goal of this talk is to explain the Bitcoin algorithm from the distributed computing perspective.
Rachid Guerraoui
EPFL