Infrastructure Design and Challenges for High-Reliability OpenStack Clusters at LY Corporation

At LY Corporation (formerly LINE Corp), we are committed to providing a highly reliable and available cloud environment through our Private Cloud service, Verda.

A key component of our strategy to achieve high availability involves constructing our cloud infrastructure across multiple Availability Zones (AZs).
This approach ensures that even if one AZ experiences a failure, the remaining AZs can continue to provide uninterrupted service.

In this session, we will delve into the specific measures taken to enhance the availability of the Verda cloud service, focusing on the infrastructure design and deployment strategies that support our high-reliability commitments.

Key topics will include:
- Multi-AZ Control Plane: How we architect the OpenStack control plane across multiple AZs to ensure continuous operation despite potential AZ failure

- Software Load Balancers: The implementation of software load balancers that facilitate redundant service connections within each AZ.

- Messaging Services: Deployment strategies for messaging services like RabbitMQ and Consul, ensuring high availability and resilience.

- Databases: Techniques for deploying critical database services such as MySQL and Redis across multiple AZs to prevent single points of failure.

Speaker(s):
- Xiang Wang

Connect with us: