Morten Ofstad, Chief Architect at Bluware
The following formats have been open-sourced to the energy industry for developers, data managers, and geoscientists to chart a path for more efficient use of volumetric data:
- VDS (Volume Data Store)
- MDIO (Multi-Dimensional Interpretation Object)
- OpenZGY
- SEG-Y
These formats can play a crucial role in the oil and gas industry, providing options for storing and analyzing vast amounts of seismic data. Let’s dive into a comparative analysis of the formats. We will explore their features, advantages, disadvantages, and use cases.
Figure 1: Commercial VDS vs Other Formats
GPU Support: GPU acceleration for processing is essential for faster data handling in compute-intensive tasks. VDS (Bluware Engine) offers GPU support, significantly enhancing performance for high-demand applications.
OSDU Compliance: VDS, MDIO, OpenZgy, and SEG-Y are accepted formats by The Open Group OSDUTM Data Platform.
Compression Capability:
VDS adaptive compression technology uses a form of Embedded Zerotree Wavelet Compression (EZW). This sophisticated approach:
- Encodes coefficients adaptively
- Functions similarly to a wavelet soft-threshold noise removal filter
- Creates smoother data at higher compression ratios, benefiting workflows like horizon interpretation
Cloud Capabilities: VDS supports cloud-native workflows, which allows teams to collaborate remotely and securely while leveraging the computational power of the cloud for AI-driven interpretation, reducing dependency on local infrastructure.
ARM 64 Support: Bluware Engine supports ARM 64 architecture, making it versatile for a broader range of deployment environments.
Adaptive Streaming Capability: Unlike other format systems that require multiple format conversions and copies of data, VDS adaptive streaming technology streams data to applications in real-time without duplicating the data, so it is available instantly. Adaptive streaming delivers precisely the signal quality needed for each workflow to dramatically improve data access speeds, enabling seamless integration into diverse geoscience processes.
Cost-Efficiency for Storage: VDS provides significant cost-saving optimizations, crucial for large datasets in cloud storage. Other formats lack similar cost-efficiency features, potentially increasing storage expenses.
Performance Optimization: Encompasses various technical enhancements for maximizing speed, efficiency, and throughput, particularly for large datasets. Bluware Engine and OpenVDS+ are optimized for:
- Efficient data retrieval through hierarchical data structuring.
- High throughput for reading and writing large seismic files.
- Support for parallel processing on multiple CPUs and GPUs (Bluware Engine only).
- Caching and prefetching strategies to anticipate data access patterns (Bluware Engine only).
- Compatibility with cloud and distributed systems, minimizing latency for remote access.
MDIO and OpenZGY have limited performance optimizations, generally lacking advanced caching, GPU support, and multi-level data access features, which restricts their speed and efficiency in intensive workflows.
Scalable Data Access: VDS can handle large data volumes and concurrent access without performance degradation, important for collaborative or high-demand environments. Other formats may struggle with scalability, causing bottlenecks as data or user load grows.
Choosing the Right Format for Your Organization
The choice on what format to use should be based on your organization’s unique needs, existing software infrastructure, and long-term goals.
VDS is highly recommended due to its adaptive streaming and compression capabilities making it ideal for both on-premises and cloud-based seismic workflows. It is used in a wide range of applications making it the ideal format for interoperability.
MDIO is optimized for multi-dimensional seismic data and efficient data access, but has limited performance optimizations, which restricts speed and efficiency in intensive workflows like machine learning.
OpenZGY provides compatibility siloed mostly with SLB’s ecosystem and the performance of large datasets is a concern.
SEG-Y, while widely used, is outdated for modern cloud and AI-driven workflows due to its limitations in performance and scalability.
Organizations should assess factors such as cloud readiness, AI integration, and computational efficiency when selecting a format.