The incentives for enterprises to pursue a multi-cloud deployment strategy—a cloud-agnostic infrastructure, greater resilience, the flexibility that comes from not being reliant on any single vendor, to name just a few—have never been more compelling, and they are constantly increasing. Yes, the technological feat of implementing and managing deployments that straddle multiple clouds comes with some challenges. But as the need for this future-ready architecture increases, Apache Cassandra is a uniquely primed open source database solution for enabling such deployments.
Here are six ways that Cassandra delivers a powerful foundation for multi-cloud architecture:
1. Topology-aware availability
Cassandra databases allow setup of failure domains, which structure replica placements around any common groups of machines with potential vulnerability to simultaneous failure (i.e., machines in a Cassandra cluster that are on the same hypervisor, rack, availability zone, etc.). In this way, Cassandra can be made aware of critical shared resources to safeguard availability, even in group outages.
2. Tunable consistency
Cassandra enables developers to control the consistency of a query in line with failure domains, within a data center or a rack. Rack-level failure domains are named after the actual physical racks within a data center, where machines share power and networking infrastructure. Cassandra can ensure that these racks are bound by high-bandwidth/low-latency connections and that replicas for each particular row are stored in separate rack locations. Doing so ensures that replica failure is consistent across the entire dataset and that all resources remain available.
3. Remote regional awareness
Data centers within Cassandra can be treated as either local or remote, so their potential to feature less bandwidth and higher latency is fully accounted for. Cassandra drivers and nodes include an awareness of the local data center, and of what other nodes are present. Utilizing Cassandra’s tunable consistency, strength of consistency can be isolated in relation to the data center where a given query is sent. Developers are thus prepared to consider how queries interact with database failure and physical topology, due to Cassandra’s ingrained awareness of remote failure domains and data centers.
4. Flexible global and local consistency
Cassandra’s fault domain model is particularly suited to the cloud, offering developers tremendous capabilities to create applications that utilize multiple data centers, availability zones, and regions. With an application that supports multiple regions in place, it isn’t difficult to extend this reach to multiple cloud providers—a secondary cloud provider’s region should feature a remote latency similar to the remote regions of the primary provider. Because Cassandra can isolate consistency to a data center, adding new regions or data centers to a cluster requires no downtime and has negligible impact. At the same time, Cassandra can guarantee developers that queries are globally dependable through higher consistency levels.
5. Simple and effective replication
Cassandra makes it easy to perform region- or data center-based data replication that’s configured at the keyspace level, so that developers can choose to replicate data to certain region subsets within a Cassandra cluster. This, in turn, gives developers a simple way to fulfill requirements such as customer and business needs or government compliance regulations. Cassandra also makes it simple to separate any applications or tables that rely on multiple cloud providers from those that relate to only one provider.
6. Open source licensing with needed flexibility
As an open source software project, any enterprise can run Cassandra with any cloud provider or private data center, retaining a level of freedom and flexibility that proprietary database solutions or cloud provider services simply cannot match. Whereas database services delivered by cloud providers are designed to promote vendor lock-in, Cassandra provides a simple and cost-effective alternative to avoiding lock-in with a single cloud vendor.
Achieving a true multi-cloud or hybrid deployment is challenging, given the differences in cloud provider environments and the need for separate process and code paths required to handle config management, CI/CD, alerting, resource planning, and routing/DNS. However, Cassandra’s strength as an open source database suited to multi-cloud deployment means that the data layer is free of these issues. Teaming Cassandra with other popular standardized solutions can empower an enterprise with an application stack that’s extendable across multiple cloud providers, which is especially well-timed as more enterprises turn toward a long-term multi-cloud strategy.
Comments are closed.