Behind every SQL table lies a carefully engineered storage system. In GBase 8s, data is not simply stored in rows—it is organized into extents, fragments, and external mappings that define how information lives on disk.
Understanding this architecture is essential for performance tuning and system design.
What Is a Storage Extent?
An extent is a contiguous block of disk space allocated to a database object.
Instead of allocating space row by row, databases allocate in chunks:
- Initial extent (base allocation)
- Next extents (growth allocations)
This approach reduces fragmentation and improves sequential read performance.
How Extents Control Growth
When a table grows beyond its initial allocation, GBase expands it using predefined rules.
These rules define:
- Initial allocation size
- Increment size for future growth
- Storage distribution behavior
This ensures predictable performance even as data volume increases.
Fragmentation and Data Distribution
Large tables are often split into fragments to improve scalability.
Fragment metadata is stored in system catalogs and used to:
- Distribute data across storage spaces
- Balance I/O load
- Improve parallel query execution
This makes GBase suitable for high-volume workloads.
External Tables: Bridging Internal and External Data
Not all data lives inside the database.
GBase supports external tables, allowing the engine to query files directly (such as CSV or structured logs).
External table definitions include:
- File format rules
- Field separators
- Data type mapping
- Error handling policies
This is especially useful for ETL pipelines and data lake integration.
Metadata-Driven Storage Design
The key idea behind GBase storage architecture is that everything is metadata-driven:
- Logical structure → system catalogs
- Physical layout → fragmentation metadata
- External integration → external table definitions
This layered design allows flexibility without sacrificing performance.
Conclusion
GBase storage architecture is more than just disk allocation—it is a structured system where metadata defines how data is stored, accessed, and scaled.
By understanding extents and external tables, you gain insight into how enterprise databases achieve both performance and flexibility under heavy workloads.
Top comments (0)