We have just launched 2 new instance types for the Distributed Cache Service (DCS) in the Open Telekom Cloud EU-DE.
Read/Write split instances
These instances are designed to improve performance in scenarios with high read concurrency and few write requests. This feature uses an automatically deployed Elastic Load Balancer to distinguish between read and write operations, forwarding write requests to the master node and read requests to the standby node(s).
Advantages:
- Improved Read performance offering high read concurrency efficiently.
- Automatic Read/Write distribution to simplifying client integration as no additional configuration is needed.
- High availability using built-in failover mechanisms.
- Higher cost efficiency by offloading read operations to replicas, the architecture reduces the load on the master node, improving resource utilization.
Proxy Cluster instances
These instances help to achieve high availability and support millions of concurrent requests. These instances offer a wide range of memory specifications, from 4 GB to 1024 GB, adapting to various scenarios and providing distributed data storage and access without requiring special client configurations. Proxy Cluster instances can be connected using the IP address or domain name, simplifying integration while offering performance comparable to DCS Cluster instances.
Advantages:
- Distributed data storage and access without requiring special client configurations.
- Ability to connect using one IP address or domain name, like single-node or master/standby instances.
- Multi-AZ deployment support for improved availability.
- Automatic failover between master and replica nodes within shards.
Both new instance types support backup functionality. A DCS Proxy cluster is preconfigured with one replica per cluster node, providing high availability out of the box. Additionally, it allows customized sharding.
More details in the Documentation.