Google Cloud unveils ultra-low latency tools for markets

Google Cloud has made its Ultra Low Latency Solution and U4 machine family generally available for capital markets users. The offering is aimed at exchanges, trading firms and service providers running latency-sensitive workloads in select private cloud regions.

The launch is Google Cloud's latest push into market infrastructure, as exchanges and trading firms look for cloud systems that can match the timing predictability of physical co-location.

The package combines dedicated bare metal and virtual machine options with networking, storage and observability tools for high-frequency trading workflows. It is intended to support market-data feeds, order routing, trade validation, packet capture and regulatory audit requirements.

At the centre of the release is the new U4 machine family, split into three series. U4P and U4C are dual-socket bare metal instances for exchange operators and market participants, while U4S is a virtual machine series for operators, participants and service providers.

The bare metal systems use Intel 5th Gen Xeon processors with 120 physical cores, 512 GB or 768 GB of memory, and three physical network interface cards. The U4S virtual machine series uses Intel 6th Gen Xeon processors and scales from 2 to 288 vCPUs, with up to 2,232 GB of memory.

The Ultra Low Latency setup is built on three main elements: hardware-level multicast distribution for market-data feeds, hardware-timed network observability, and predictable processing for execution workloads. It also uses separate network adapters for market feeds and for telemetry and management traffic.

Operational features include direct-attached Titanium SSD storage for real-time logging, Hyperdisk storage for archives and analytics, and support for OpenOnload and DPDK for packet processing in Linux user space. The system also integrates with Google Cloud's Firefly clock synchronisation technology to provide sub-10 nanosecond network-level timestamping and UTC synchronisation, beyond the sub-100 microsecond regulatory threshold cited for exchanges.

Market focus

The launch comes as cloud providers seek a bigger role in financial market infrastructure, an area long dominated by on-premises hardware in exchange data centres. Low latency and stable packet timing are central to trading performance, especially in listed derivatives and other markets where queue position can affect execution.

Google Cloud pointed to its work with CME Group as an example of how the system is being used. The companies have been working on the migration of listed derivatives markets to Google Cloud, with the new bare metal instances intended to support exchange and participant workloads that have traditionally remained in physical co-location facilities.

"Migrating the world's leading derivatives marketplace to the cloud requires uncompromising performance, latency, and reliability. Google Cloud's Ultra Low Latency Solution and the U4 instance families represent a breakthrough for financial market infrastructure. By combining deterministic, ultra-low latency networking and dedicated bare metal compute with the benefits of the cloud, market participants can execute and ingest market data with the speed of traditional co-location and the agility of the cloud," said Pearce Peck-Walden, Managing Director of Markets Engineering at CME Group.

Testing results

Other market firms cited by Google Cloud said they had tested the systems for trading use cases. Prime Trading said the combination of bare metal and virtual machine options allowed it to compare performance and cost across different strategies.

"Our testing of the U4 instances validated that Google Cloud delivers the low latency and determinism required for our demanding exchange trading workloads. Having both bare metal and VM options allows us to evaluate the value of that performance relative to its cost, helping us align the right infrastructure to each of our trading strategies," said Corbin Kidd, EVP and CTO at Prime Trading.

28Stone said its tests focused on tick-to-trade performance and consistency under stress. Determinism, or the ability to keep performance flat under load, remains a core requirement for institutional trading systems.

"In our tick-to-trade performance testing, the U4C bare metal instance demonstrated exceptional speed, showing that cloud compute and ultra-low latency networking can comfortably satisfy microsecond trading demands. But what sets the U4 series apart isn't just raw processing speed, it's structural determinism. When we stress-tested Google Cloud's U4 bare metal family, the performance curve remained virtually flat, confirming that these cloud compute layouts can now offer the strict determinism institutional market participants require," said Christopher Wilson, Global Head of Experience Modernization at 28Stone.

The Ultra Low Latency Solution is available in select private regions and can be managed through standard cloud APIs, infrastructure-as-code tools and security policies. Customers can also use compact placement policies to keep related machines close together in the data centre, or spread placement policies to place systems across separate racks for resilience.