Engineering the Future of Software
18-19 October 2016: Training
19-21 October 2016: Tutorials & Conference
London, UK

Schedule: Distributed systems sessions

10:45–12:15 Thursday, 20/10/2016
Location: Windsor Suite Level: Intermediate
Jakub Korab (Ameliant)
Average rating: ***..
(3.33, 3 ratings)
“Messages will be processed once and only once, and in order.” Such statements show a deep misunderstanding of messaging, a technology that is at the core of so many of the systems we build. Jakub Korab outlines the rationale and mechanics of ActiveMQ and Kafka, explains how their guarantees and constraints impact your designs, and explores how to reason about similar systems in future. Read more.
13:15–14:05 Thursday, 20/10/2016
Location: Park Suite (St. James/Regents) Level: Advanced
Wes Chow (Cortico at MIT Media Lab)
Average rating: ****.
(4.40, 5 ratings)
Want to scale? Shard it! If only it were that simple. There are more gotchas than you might expect, like possibly taking down your entire system and guaranteeing failure. Wes Chow covers the math of engineering better failure scenarios and real-world applications. Read more.
14:15–15:05 Thursday, 20/10/2016
Location: Park Suite (St. James/Regents) Level: Beginner
Kai Wähner (Confluent)
Average rating: **...
(2.50, 4 ratings)
IT applications and microservices generate terabytes of distributed log data from mobile devices, the Internet of Things, and social networks. Kai Wähner compares open source frameworks and SaaS cloud solutions for operational intelligence and log analytics, explains how these differ from big data stores such as Hadoop, and offers a live demo demonstrating how to analyze distributed microservices. Read more.
10:45–12:15 Friday, 21/10/2016
Location: Park Suite (St. James/Regents) Level: Intermediate
Tim Berglund (Confluent)
Average rating: ****.
(4.50, 2 ratings)
Tim Berglund explores four ready-to-use examples of how to architect a distributed data-processing system using open source tools, including criteria for how to select each one. Each reference architecture is based on a successful real-world production system. Read more.
13:15–14:05 Friday, 21/10/2016
Location: Windsor Suite Level: Intermediate
Kaz Sato (Google)
Average rating: ***..
(3.60, 5 ratings)
The largest challenge for deep learning is scalability. Google has built a large-scale neural network in the cloud and is now sharing the power. Kazunori Sato introduces pretrained ML services, such as the Cloud Vision API and the Speech API, and explores how TensorFlow and Cloud Machine Learning can accelerate custom model training 10x–40x with Google's distributed training infrastructure. Read more.
13:15–14:05 Friday, 21/10/2016
Location: Park Suite (St. James/Regents) Level: Intermediate
Rotem Hermon (SAP)
Average rating: ***..
(3.50, 2 ratings)
Rotem Hermon introduces the concept of “virtual actors”—a new abstraction on top of the actor model that is designed to be cloud native and aims to accelerate the development of distributed applications. Rotem covers the classic actor model, why virtual actors make distributed application programming a lot simpler, and which open source implementations are available for .NET and Java. Read more.
13:15–14:05 Friday, 21/10/2016
Location: Victoria (A/B) Level: Intermediate
Ted Malaska (Capital One)
Average rating: **...
(2.29, 7 ratings)
Distributed execution is like single-node execution, but everything is faster, right? Well, the reality is that distributed execution is complex; you must have a strong understanding of the architecture fundamentals to be successful. To help you out, Ted Malaska explores the five biggest architecture mistakes of moving to distributed computing. Read more.
15:50–16:40 Friday, 21/10/2016
Location: Park Suite (St. James/Regents) Level: Intermediate
Kristoffer Dyrkorn (Norwegian Broadcasting Corporation)
Average rating: ****.
(4.00, 2 ratings)
Kristoffer Dyrkorn outlines an infrastructure that reduces the cost of road construction and maintenance. The system provides high-quality and near real-time information by integrating an unusual combination of sensors, devices, protocols, and software. Kristoffer discusses the architectural challenges encountered and choices and mistakes made trying to keep the architecture receptive to change. Read more.
16:50–17:40 Friday, 21/10/2016
Location: Blenheim Room Level: Beginner
Josh Long (Pivotal)
Speed of evolution is a differentiator, but velocity for velocity’s sake is dangerous. Microservices invite architectural complexity that few are prepared to address. Josh Long explores how high-performance organizations like Ticketmaster, Alibaba, and Netflix make short work of that complexity with Spring Boot and Spring Cloud. Read more.