Latency: The Race to Zero...Are We There Yet?

Low and predictable latency have been an edge in financial trading. Aeron has been pushing the limit on what is possible for IPC, on-premise, and in the cloud messaging. Can we do better? Reaching back into computer science history, techniques like Virtual Synchrony can be used to, not only reduce latency, but also provide high-availability and avoid the complexity of distributed concurrency. In this talk we will explore an approach for system composition that some financial organizations use to give an edge when building mission critical systems.

Interview:

What is the focus of your work?

I work on a future product called Aeron Sequencer at Adaptive.

It is based on the cluster available in the open source Aeron project that allows teams to design and build trading systems with low and predictable latency.

What’s the motivation for your talk?

Exploring ideas at the heart of distributed systems such as Virtual synchrony, Sequencer and state machine composition. Also asking "why?" quite a lot.

Who is your talk for?

Developers, architects, anyone who likes to ask "why?", take a wider view and think about system design as part of the solution, not just a constraint.

What do you want someone to walk away with from your presentation?

Understand how (and why) key concepts from distributed systems design can be used to improve latency metrics.

What do you think is the next big disruption in software?

Yes, I also think the next big disruption is AI but not in the same dark scary way it is often described.

The ability for AI to do the simple, "solved", boring stuff well, means we will be forced to do a much better job of separating concerns and domain design.

We will need to design more modular solutions and do better at protocol and API design, in order to use AI to our advantage.


Speaker

Amir Langer

Principal Software Engineer @Adaptive Financial Consulting

Amir has been enjoying the challenges of software development for the past 30 years.

His career started with an MSc in computer science, then moved to software engineering roles (using Java since its first release), focusing on large scale distributed and high-performance low latency systems. In the finance domain, Amir was the first developer at LMAX exchange and a team lead for many years. Now he is a principal developer at Adaptive, the home of Aeron, collaborating on future Aeron products.

Read more
Find Amir Langer at:

Date

Wednesday Apr 9 / 03:55PM BST ( 50 minutes )

Location

Churchill (Ground Fl.)

Topics

architecture aeron performance sequencer

Share

From the same track

Session

Fighting Financial Crime with AI

Wednesday Apr 9 / 01:35PM BST

Details coming soon.

Session Platform Engineering

Unleashing Kubernetes for Secure Bare-Metal Workloads

Wednesday Apr 9 / 11:45AM BST

Kubernetes is great for general cloud-native workloads but struggles with low latency and high-performance computing (HPC) due to its abstraction overhead, lack of optimized scheduling, and network inefficiencies.

Speaker image - Ruslan Kharitonov

Ruslan Kharitonov

Head of Compute @Citadel, Founder @Efacity, Previously Technology Fellow and Global Head of Network and Storage Software Engineering @GoldmanSachs, 25+ Years in Software and System Engineering

Session database

How to Build a Database Without a Server

Wednesday Apr 9 / 10:35AM BST

Modern data analytics workflows rely on scaling out to huge numbers of users and compute nodes. Managing database installations to handle this scale can be unsustainably complex and expensive.

Speaker image - Alex Seaton

Alex Seaton

Staff Engineer @ArcticDB, Previously Working on Quant Trading Systems @Man Group

Session

Navigating the Regulatory Landscape in FinTech

Wednesday Apr 9 / 02:45PM BST

Details coming soon.

Speaker image - Sérgio Amorim

Sérgio Amorim

DevOps Engineer @Revolut, DevOps Lisbon Meetup Organizer