- Products
- Solutions
- Resources
- Company
- Pricing
- Contact Us
Start Free
Challenges
High-concurrency reporting and analytics services with a large number of real-time user access, including marketing reports (ad exposure, clicks, and spending), insurance analysis (insurance plan tailoring and customer conversion reporting), logistics dashboards (real-time analysis of logistics pressure, efficiency and customer complaints), and transaction reports (order, bill, and delivery queries).
The above shows the common online reporting solutions. With high-concurrency user access, these solutions are often challenged by the following problems:
High data latency
Online reporting requires the latest data and a data latency within seconds, while the common solutions only support batch updates without reflecting the real-time changes of data.Slow query response
Sub-second query latency lays the foundation for good user experience, but common solutions often struggle to guarantee quick query response with large data sizes.Low query concurrency
Online reports serve a large user base that includes not only management and analysts but also general customers. Common solutions often find it hard to handle high-concurrency requests.Low service availability
Service failure of online reports in logistics and retail is detrimental to business. Unfortunately, this is often the case with common solutions under high concurrency or big query circumstances.
The VeloDB solution
Real-time, stable, and reliable services in high-concurrency reporting
Real-time data Ingestion- ·A writing throughput of 1 million rows per second. Support streaming data from OLTP databases and Kafka.
- ·Support real-time data updates, such as order status, to generate real-time reports.
Sub-second query response- ·Full vectorization from storage to computation to increase query speed by orders of magnitudes.
- ·Strongly consistent materialized views for pre-aggregation to deliver sub-second response.
High-concurrency support- ·Queries are distributed across machines for execution. Reduce data scanning and increase query concurrency by indexes like ZoneMap, BloomFilter, and inverted index.
- ·Reduce IOPS pressure by row cache to support 10,000 QPS per node.
High service stability & availability- ·Single-cluster availability: online cluster scaling and auto-balancing of replicas, no single point of failure.
- ·Cross-cluster availability: cross-cluster replication in real time.
Try VeloDB now
VeloDB Cloud
Fully managed, cloud-native, real-time data warehouse service
Start free
VeloDB Enterprise
Self-managed software on premises, on VMs, or K8s
Start free