Registration for our Summer Community MeetUps is now open. Secure your slot. Register now →
The CME-Google Cloud acceleration: why market makers need to speed up cloud migration and how to bridge the cloud performance gap
The “Cloud Migration” timeline for liquidity providers just suffered a compression. Following the latest rumours from FIA Boca, the industry’s assumption of a comfortable lead time for the CME’s move to the cloud has been upended.
CME is moving some markets to the new Google Cloud Dallas region this year. While Dallas serves as the primary location for client validation and testing (sandbox), its role as the future disaster recovery (DR) site—and the precursor to the Chicago production migration—means the clock is now ticking. Client testing in the Dallas sandbox began in early 2026, and production migration is slated to start within the same year.
What was once a multi-year evaluation phase has become a multi-month execution mandate. For Tier 1 market makers, the move to a cloud-native environment is hence an immediate operational requirement.
The hurdle: standard cloud networking meets markets reality
For liquidity providers, the cloud presents a fundamental problem: standard cloud networking—relying on GVE (Google Virtual Ethernet) and traditional TCP stacks—is designed for throughput and general-purpose resilience, not for the demands of market participants.
On-premises, firms build their edge on bare metal: bypassing the kernel, pinning threads to isolated cores, controlling every inch of the NIC, and working in FPGA. In a virtualized cloud environment, trading firms lose this control.
While Google has made significant, market-grade investments in low-latency infrastructure (C3/C4 instances and Titanium offload engines), the cloud remains a dynamic, shared-resource environment. In an industry where every microsecond of jitter can translate to slippage or missed fills, the software layer must be as specialized as the hardware to ensure determinism, performance or high availability.
Aeron is the industry-standard solution for firms that cannot afford to compromise on performance while making their transition to cloud trading. It was built from the ground up to be cloud-tolerant, providing a high-performance messaging and clustering framework that addresses the complexities of cloud infrastructure.
Performance: Low-latency, high-throughput data transport and dissemination
High-availability: Resilient, performant, stateful systems
By deploying Aeron, market makers can approach on-premises performance within a cloud environment, maintaining a deterministic execution environment in a shared-resource world.
The core Aeron value proposition for your cloud journey
01 DPDK Kernel Bypass (Aeron Premium)
High-frequency execution in the cloud is often hamstrung by the standard networking stack. Leveraging DPDK is the engineering solution; however, building a custom DPDK-accelerated stack from scratch is a significant engineering undertaking—one that most firms cannot afford given the new compressed CME timelines.
Aeron Premium provides an out-of-the-box DPDK (Data Plane Development Kit) accelerated stack. This allows for kernel bypass in the cloud, significantly reducing CPU overhead and eliminating the latency inherent in standard TCP/IP stacks.
02 Proven in production
Aeron’s cloud tolerant capabilities aren’t theoretical. Aeron is already the backbone for major cloud-native exchanges and trading venues. Bullish is only one high-profile example of an institution that has successfully leveraged Aeron to maintain high-throughput, low-latency operations in production environments.
Don’t just take our word for it. Aeron’s performance is backed by reproducible benchmarks conducted on both Google Cloud and AWS. Core Google Cloud Findings:
Aeron Transport comes together with Google Cloud’s latest developments in C3 instance types to allow the open-source edition to achieve round trip times of sub-60 microseconds. When combined with DPDK in Aeron Premium, Aeron Transport is able to transmit data reliably and at sub-20 microsecond rates on Google Cloud and achieve throughput of around 5 million messages per second.
Using Aeron Cluster and Google Cloud together, organisations can achieve millisecond-level recovery with zero data loss while maintaining latency of around 100 microseconds and throughput of multiple million messages per second. Aeron Premium improves the latency of Aeron Cluster by 3 fold and throughput by almost 9 fold.
De-risking the capital markets cloud migration journey
The move to the cloud shouldn’t be a leap of faith. It should be a calculated migration to a more scalable architecture. Aeron provides the safety net, ensuring that as you move your liquidity provision into Google Cloud’s Dallas or Chicago environments, your execution remains as sharp as it was in the local data center.
The timeline might have accelerated, the technical requirements haven’t changed.
