Application Scenarios
AI-oriented vector storage, search, and analysis
With the advancement of AI technology, particularly the emergence of large language models, unstructured data analysis has become both feasible and easy to employ. Relyt offers hybrid search and analysis capabilities for both structured and unstructured data, enabling you to process various data types such as video, audio, images, and text through a simple and easy-to-use SQL interface.
Surging demands in queries and analytics
In many business scenarios, data processing workloads exhibit significant variability, characterized by frequent peaks and troughs. As a result, scaling compute resources dynamically is essential to handle unpredictable spikes in workload efficiently. Traditional data warehouse solutions often lack real-time elastic scaling capabilities, although some offer limited elasticity. This often leads to over-subscription of compute resources, sometimes reaching as high as 80%, thereby increasing the total cost of ownership (TCO) for organizations.
Relyt's DPS clusters support a wide range of types and specifications tailored to diverse customer needs. These clusters can be deployed, configured, resumed/suspended, and scaled automatically on demand, facilitating real-time expansion and reduction of compute resources. Additionally, DPS clusters' feature AQS, which automatically identifies and directs large queries during execution to a shared computing resource pool. This ensures stable and uninterrupted business operations while managing costs effectively.
Reduce storage costs for analytics systems
In traditional data processing scenarios, storage often becomes a major bottleneck for organizations, second only to computing. Different compute engines and systems may require multiple redundant data copies, increasing storage costs and complicating data management. This issue is particularly prominent in data-intensive industries such as gaming, SaaS, finance, e-commerce, and IoT.
Relyt's vectorized engine, Vector DPS, supports a hybrid storage model, allowing Relyt to maintain a single global copy of data without redundant replicas. Additionally, Relyt's innovative features, such as sort key-based pruning, significantly reduce system I/O operations by up to 99%. Moreover, with its hybrid storage model, Relyt supports adaptive selection of compression algorithms and encoding methods, ensuring high I/O efficiency and minimizing storage costs.
Simplify data pipelines
Data processing is inherently complex, and the intricacy of data pipelines is often one of the most significant challenges. Typically, data must pass through multiple systems and ETL tools, which not only complicates workflows but also significantly increases operational difficulty and costs. This is especially problematic when trial and error is involved, as complex pipelines can make experimentation extremely costly.
Relyt AI-ready Data Cloud supports various data formats, including structured and semi-structured data, and can handle diverse workloads such as analytical, transactional, and data science & AI tasks. With Relyt AI-ready Data Cloud, core business needs such as data storage, querying, and analysis can be addressed in a one-stop solution without the need for ETL, minimizing processing complexity while ensuring real-time data availability.
Require high system agility
As businesses expand, data accumulation continues, and many organizations find their existing data warehouse systems constrained by resource and performance bottlenecks. These systems struggle to meet the demands of data storage, computing, and analysis brought on by business growth. There is an urgent need for a system that can swiftly and agilely adjust to business changes, ensuring stability in data analysis. For instance, new SQL queries might impact existing operations, leading to a reluctance to use the system.
Relyt offers exceptional real-time performance, elasticity, and reliability with its "storage-compute separation + SHARED-DATA" architecture. Relyt's DPS clusters support on-demand deployment and scaling, flexibly meeting the agile requirements of evolving business needs.
Require highly stable analytics systems
As your organization develops, the volume of data grows exponentially. SQL queries to address new business needs may need to scan large amounts of data, consuming significant compute resources of the data warehouse system. If system resources are improperly allocated or processing capacity is insufficient, this may affect online operations, leading to system instability. Allocating sufficient resources often results in high resource costs.
Furthermore, as business concurrency increases, data write and query concurrency will also rise accordingly. This will likewise consume substantial computational resources of the data warehouse system. If the system cannot effectively handle high concurrent requests, it may impact online operations, causing system instability. Therefore, the data warehouse system needs to possess high concurrency processing capabilities to maintain stable operation under high concurrency conditions.
