In today's data-driven world, data unification is crucial to ensuring data accuracy and consistency across an organization. Reltio, a leading provider of AI-powered data unification and management solutions, recently took a significant step in modernizing its data architecture by moving from self-managed Cassandra to Spanner on Google Cloud. The outcome of this work shows improved performance capabilities and seamless scalability.
Reltio provides AI-powered data unification and management solutions that provide unified, reliable data where and when it's needed, allowing businesses to adjust to changing business requirements. The Reltio Connected Data Platform powers Reltio's products, which include 360 Data Products, Multidomain Master Data Management (MDM), and Entity Resolution. These tools create a single source of truth for decision-making by combining multiple data sources in milliseconds. Reltio offers real-time, high-quality data that is AI-ready and has scalability, security, and flexibility. Leading businesses in a variety of sectors benefit from Reltio's ability to boost productivity, lower risk, and accelerate development.Because of these heavy workloads, Reltio needed a database that is scalable, has low latency, and provides the reliability that the Reltio Connected Data Platform demands. Furthermore, migration must be easy. Spanner was chosen by Reltio since it satisfied all of its requirements.
The Reasons Spanner Was the Best Option
- Smooth Scalability and Performance: Spanner's straightforward scalability and seamless workload adaptation allowed Reltio to respond to shifting needs and always maintain peak performance.
- Better Availability and Reliability: Compared to self-managed Cassandra, Spanner has built-in reliability upgrades that provide peace of mind, along with high availability guarantees (99.99% for regional and 99.999% for multi-regional).
- Unlocking Innovation: Spanner's broad feature set, which provides integrated search capabilities and seamless integration with other Google Cloud services, allows Reltio to explore new features and enhance their data unification offerings. ACID transactions provide unparalleled data consistency, and multi-regional instances enable its services to be more broadly accessible.
- Streamlined Operations: Because Spanner is completely managed, there is no need to worry about infrastructure maintenance, which leads to significant cost savings and streamlined operations.
Reltio's Spanner Architecture and Migration Plan
Let's look at the technical architecture and migration strategy that enabled Reltio to seamlessly transition to Cloud Spanner.Tenant isolation and security
- By keeping each tenant's data in a different Spanner database, total physical data plane isolation may be guaranteed.
- Spanner's robust security design tightly controls access to each database.
- Compute plane uses a service account with permissions to temporarily assume roles in order to facilitate data access.
Autoscaling and Capacity Management
- A hybrid autoscaling paradigm uses both human scaling triggered by client-side latency criteria and automated scaling based on CPU and storage metrics.
- This approach ensures optimal performance and resource usage while guarding against over-scaling caused by inadequately throttled tenants.
Backup, Restore, and Disaster Recovery
- Point-in-time recovery (PITR) backups are implemented using database export/import and are thereafter saved in Google Cloud Storage with a specific lifespan setting for cost-effectiveness.
- Recuperation from local catastrophes Backups are stored in secondary Spanner instances located in backup regions, allowing for rapid recovery in the case of a regional outage.
Tenant Provisioning and Monitoring
- A cloud function automates instance assignment, tenant provisioning, and service account creation.
- Another cloud function monitors tenant consumption and instances to optimize autoscaling settings and remove wasteful resources.
Preventing Throttling and Noisy Neighbors
- Spanner modifies Reltio's internal traffic-shaping, which is based on credits, to ensure that tenants in shared instances receive a fair distribution of resources.
- There is an integrated Spanner autoscaler function that looks at per-database metrics to prevent over-provisioning and protect against workloads from one tenancy impacting others.
- This is an excellent example of resource management at the database and application layers working together to deliver a consistent user experience, and it is a best practice for a well-architected program.
0 Comments