Upfront of Ignite, I spoke to Shireesh Thota, CVP Databases at Microsoft, in regards to the new service. He described the rationale for a brand new PostgreSQL variant:
I feel more and more what we discover is that individuals both go into the bucket of, “I need to raise and shift my PostgreSQL that’s working locally model on-premises, or possibly one other cloud.” They need to transfer it to Azure. They need 100% Postgres. They need all extensions working. They simply need one thing that actually has the pliability of efficiency and velocity. Then Azure Database for PostgreSQL, the present model is ideal. Someone who desires to construct an AI-native, cloud-native form of a workload which will want quite a lot of storage, desires actually quick latencies, considerably greater IOPS. Then you definately go to HorizonDB.
Definitely, the printed efficiency information for Azure HorizonDB is spectacular: Microsoft is claiming a three-times improve in throughput over the open supply launch when working transactional workloads. You may scale as much as 3072 cores, with 128TB of storage and sub-millisecond commits. HorizonDB builds on Azure’s multiregion structure with information replicated in a number of availability zones and automatic upkeep and backups with minimal influence on operations. Such efficiency is required for AI purposes and for large-scale Kubernetes. As Thota notes, “These cloud-native workloads can actually succeed on HorizonDB.”
Key to the efficiency increase are modifications to the structure of the database, separating compute and storage and permitting them to scale independently. If you happen to want extra compute, Horizon DB will give it to you. If you happen to want extra learn replicas, it’ll provision them.
