Scenarios
LHC is a service for cloud-native environments. It is a Platform as a Service (PaaS) offering that unifies the management of applications and resources across multiple clusters. The platform provides a centralized view for releases and O&M, and enables cross-cluster deployment, resource management, and traffic management.
This includes the following scenarios:
Intra-city active-active
This scenario uses multiple Kubernetes clusters across two or more zones in the same region.
Three data centers across two regions
This architecture builds on the intra-city active-active model. It adds a remote data center for data and application backups. Three remote backup options are available depending on network latency and bandwidth: hot, warm, and cold standby.
You can choose hot standby if the latency from the remote location is acceptable for your services. In normal operation, traffic is directed to the intra-city active-active deployment. During a disaster, traffic fails over to the remote geo-disaster recovery data center to ensure service availability.
Active geo-redundancy (multi-region active-active)
The data layer is partitioned using sharding. Different Availability Zones (AZs) can be divided into logical units, called Logical Data Centers, to process different data shards. Data access paths from the access layer to the application layer and then to the data layer should avoid cross-zone calls. This architecture allows for an active-active deployment across any number of regions.
Hybrid cloud on heterogeneous infrastructure
Kubernetes abstracts the differences in the underlying Infrastructure as a Service (IaaS). This allows services to be deployed on both Apsara Stack and the public cloud and managed with unified O&M. This model also takes full advantage of public cloud resources. In this scenario, finance customers can achieve the following goals:
Reduce investment in development and testing resources: Deploy production applications on Apsara Stack and deploy development and testing applications on the public cloud as needed.
Meet rapid on-premises disaster recovery needs: Deploy an on-premises environment to meet regulatory requirements for disaster recovery. This environment handles unexpected public cloud events (for example, Tianhong Yu'e Bao).
Elastic scale-out: Combined with an active geo-redundancy architecture, services can scale out horizontally and infinitely at the data center level as needed.