Solution
Deutsche Börse Group migrated mission-critical post-trade and DAX index systems to Google Cloud, replacing on-premises infrastructure with a scalable, latency-sensitive architecture. Financial transaction data flows through Compute Engine instances (N/C series machines) running settlement and index calculation workloads, backed by Hyperdisk Balanced and Extreme for consistent millisecond-level database performance. BigQuery ingests market indices and transaction analytics. The infrastructure supports 13,000+ daily DAX index calculations, 30M ticks, processing over 1M transactions daily while maintaining regulatory compliance and enabling provisioning times to drop from months to hours.
Data flow
Trading and settlement transactions ingest into Compute Engine instances where they are processed in real-time. Core databases on Hyperdisk Extreme store transactional state with millisecond persistence. Index calculation engines on C4D machines compute 13,000+ indices and 30M daily ticks, with aggregation results flowing to BigQuery for analytics and regulatory reporting. Microsecond-level latency between components ensures sub-50ms settlement turnaround.
Solution architecture
6 components · 4 layers - Compute
- Compute Engine Runs latency-sensitive transaction processing and settlement workloads using C4D compute-optimized and N4 general-purpose instances for financial applications.
- Storage
- Hyperdisk Balanced Provides consistent storage performance for post-trade processing databases handling millions of daily financial transactions.
- Hyperdisk Extreme Delivers high-performance storage for core DAX index calculation and trading databases requiring sub-millisecond latency.
- Serving
- BigQuery Aggregates index calculation results, market data analytics, and business intelligence for the DAX family and EUROSTOXX50 indices.
- Orchestration
- Google Kubernetes Engine Orchestrates microservices-based applications for new financial services, enabling scalability and continuous delivery across business lines.
- OpenShift on Compute Engine Maintains portability and standardization of containerized applications using Intel-based CPUs on Google Cloud infrastructure.
- Clearstream core financial apps run on Compute Engine N-series and latest C-series; latency-sensitive databases on C4D; broader app services on C4/N4
- Cortex, BigQuery and AI used downstream to accelerate innovation and reduce cost
- Customer-managed encryption keys, zone separation and regional failover safeguards as standard compliance controls
- DAX family and EUROSTOXX50 index calculations executed in the cloud — 13,000+ indices, ~30M ticks/day
- Dual strategy in progress: migrate traditional enterprise systems to Compute Engine, build new microservices on Compute Engine and GKE
- Hyperdisk Balanced and Hyperdisk Extreme underpin core database workloads
- Multi-zone highly resilient architecture engineered for zero data loss while meeting regulatory requirements
- OpenShift clusters on Compute Engine using Intel-based CPUs to preserve portability while standardizing on Google Cloud
- SAP S/4HANA migrated from on-premises to a secure Google Cloud foundation
33% lower TCO
50% less latency for DAX® index engine: 13k+ indices, 30M ticks
85% reduction in disaster recovery time for critical SAP S/4HANA systems
Clearstream engineered a multi-zone, highly resilient architecture on Google Cloud to achieve zero data loss for workloads while meeting strict regulatory requirements.
Latency-sensitive databases run on high-performance C4D machines optimized for compute-intensive transaction processing, while C4/N4 instances support broader application services.
Settlement times improved up to 50% for greater efficiency
