GBase Database High Availability Solutions Ensure Data Integrity-Relentless Pursuit of Business Continuity

“No data loss, no business downtime” — this is the core requirement for critical business systems. It also sets a high bar for the high availability (HA) capabilities of transactional databases. To effectively serve key business scenarios, databases must deliver robust HA solutions to ensure maximum business continuity.
But what is the mainstream database HA solutions today?
How should enterprises choose the most suitable HA solution?
And how can they better ensure uninterrupted operations?

To address these questions, the General Manager of the GBase 8s (GBase 8s 数据库) Product Management Department, accepted an interview with the database technology community ITPUB recently: “The pursuit of business continuity knows no bounds. The current database market offers a wide range of HA architectures, and enterprises need to select solutions based on their specific business needs.”

The Evolution of Database High Availability Architectures

High availability (HA) refers to a system’s ability to continue operating and providing services in the face of various failures, such as hardware malfunctions, software crashes, or network issues. This is achieved by minimizing downtime through redundancy, automated failover, and the elimination of single points of failure, ensuring stability and continuity.

In the early days, databases lacked HA architectures. Instead, availability was ensured through regular backups and log archiving, leaving databases vulnerable to single points of failure. HA at the time was primarily managed at the operating system level, using clustering software like IBM HACMP for AIX systems to provide failover functionality.

The 1990s saw a surge in internet adoption and growing reliance on data, making HA a critical design factor for databases. Redundancy and failover mechanisms became standard, enabling higher reliability. Over time, HA architectures evolved alongside advancements in storage, networking, and business needs, moving from standalone setups to master-standby clusters, shared storage clusters, and distributed clusters.

For example, Oracle, a leader in the industry, has introduced various HA solutions over the years, including Highly Available Hosts (HAH), Active Data Guard (ADG), Real Application Clusters (RAC), and Oracle GoldenGate (OGG). These solutions, either standalone or combined, have provided robust HA services for mission-critical applications in industries like finance and telecommunications.

• ADG (Active Data Guard): Initially called Data Guard (DG), ADG in Oracle 11g allowed standby nodes to support queries, reducing load on the primary node and enhancing HA. Oracle later offered three modes — maximum protection, maximum availability, and maximum performance — to provide flexible data protection options.
• RAC (Real Application Clusters): Introduced in Oracle 9i (2000), RAC is a shared-everything architecture enabling multiple servers to share a single database instance. This provides HA and load balancing, ensuring continuity even if a node fails.

GBase 8s: A Complete HA Solution for Business Continuity

Many database vendors, inspired by Oracle’s approach, have developed their own comprehensive HA stacks. For example, GBase 8s offers a full suite of HA technologies, including:

1) Master-Standby Clusters:

• HAC (High Availability Cluster): A one-master-one-standby architecture supporting synchronous, near-synchronous, and asynchronous redo log replication, mirroring Oracle ADG’s three modes.
• RHAC (Remote HAC): Supporting one-master-multiple-standby setups with asynchronous replication, suited for remote disaster recovery.

2) Shared Storage Cluster (SSC):

• Similar to Oracle RAC, SSC enables HA via shared disk storage, ensuring data consistency and supporting multi-write and multi-read capabilities. This setup effectively utilizes hardware resources and avoids redundant data storage.

3) Real-Time Data Synchronization (ER):

• A solution akin to Oracle GoldenGate (OGG), enabling real-time data synchronization at the table level, ideal for scenarios like intercity or interdepartmental data sharing.

An enterprise can deploy active/standby HA or shared storage cluster HA independently or in combination to build higher-level HA solutions such as SSC+HAC (same city with two centres) and SSC+HAC+RHAC (remote locations with three centres).

Choosing the Right HA Solution

While advanced HA architectures offer impressive capabilities, they are costly. Even well-funded institutions, such as banks, must carefully weigh their options due to the high expense of HA deployments. Factors such as data volume, consistency requirements, and budget play a pivotal role in determining the optimal solution.

• Master-Standby Clusters: Best for scenarios with smaller data volumes and moderate consistency requirements. However, they involve higher storage costs due to redundancy and may experience delays in data synchronization and failover.
• Shared Storage Clusters: Ideal for high-data-volume scenarios requiring strong data consistency. They provide a cost-effective solution by minimizing hardware requirements and ensuring continuity. However, they demand higher technical expertise and robust infrastructure, such as high-bandwidth networks.

Real-World Success Stories

Currently, the GBase 8s HA cluster solution can achieve RPO=0 and RTO < 30s. In real service scenarios, the active/standby switchover can be completed in 10–15s.

GBase 8s HA solutions have been widely adopted across industries like finance, transportation, energy, and government. Some notable implementations include:

• State Grid Corporation of China: Utilizing SSC+RHAC for a 50TB business data platform, achieving sub-second data synchronization across thousands of kilometers with uninterrupted service for over 600 days.
• Shenzhen Metro: Employing SSC+HAC for real-time ticketing and operational management, supporting over 4,000 concurrent sessions with millisecond-level response times.
• A Regional Commercial Bank: Replacing Oracle with SSC+HAC for its core system, maintaining uninterrupted operation for over 760 days.

Conclusion: The Endless Quest for Business Continuity

The pursuit of uninterrupted business operations drives continuous innovation in HA solutions.

GBase 8s is committed to refining its offerings, such as granular resource control, session persistence, and shard-level data distribution, to ensure seamless customer experiences.

As the General Manager of GBase 8s summarized: “Our ultimate goal is to make database failures invisible to customers. Achieving business continuity requires collaboration between vendors and users, with realistic HA targets and solutions tailored to real-world needs.”

In today’s digital economy, ensuring business continuity is an unending journey, but with thoughtful planning and innovative solutions, enterprises can stay resilient and competitive.