70%+ Performance Leap, Ecosystem-Driven Market Access: Apache Doris Redefines Analytical Database Capabilities

As enterprises advance their digital transformation efforts, VeloDB—the commercial cloud data warehouse powered by Apache Doris—serves as the data foundation for a wide range of core business scenarios, equipping thousands of enterprises worldwide with sub-second analytical query capabilities.:

• Real-Time Business Intelligence & Operational Decision-Making: Clients across e-commerce, gaming, and financial services use VeloDB's sub-second query engine to power live dashboards, operational scorecards, and anomaly detection systems, enabling them to move from insight to action in seconds.
• User Growth & Targeted Marketing: Leveraging real-time behavioral analytics and user profile tagging, VeloDB helps businesses deliver personalized recommendations at scale, A/B testing, ad attribution, and other growth-focused use cases.
• Logging & Observability: As a cost-effective alternative to traditional ELK stacks, VeloDB provides real-time log ingestion and interactive search capabilities at scale, meeting the needs of DevOps and security operations teams.
• Agentic AI Data Service Layer: As enterprises accelerate the adoption of AI agents, real-time data access has become a critical component of the agentic decision-making process. VeloDB provides AI agents with low-latency, high-concurrency access to structured data, supporting Retrieval-Augmented Generation (RAG), tool-use queries, and real-time context injection for multi-turn conversations, making it an essential piece of the Agentic AI infrastructure stack.

Apache Doris is widely used for real-time reporting, ad-hoc queries, unified data warehousing, log analytics, user profiling, lakehouse query acceleration, and data-as-a-service workloads. Its core capabilities are powered by an OLAP engine that combines parallel execution, vectorized computation, and pipeline scheduling.

As enterprise data volumes continue to grow, demands on underlying compute resources are increasing across every dimension—from larger datasets and higher concurrency to more stringent latency requirements. Real-time analytical systems face mounting pressure from increasingly complex SQL, surging query concurrency, denser write throughput, more intensive multi-table join workloads, and emerging lakehouse- and AI-driven processing patterns.

For analytical databases, processor performance plays a critical role in determining how efficiently the system can scan, filter, aggregate, join, sort, and exchange data. Key factors include per-core throughput, multi-core scalability, cache architecture, memory access efficiency, and network bandwidth. To further enhance performance, Apache Doris has been optimized for Amazon Graviton5. The hardware strengths of this next-gen ARM processor align closely with Doris's vectorized, parallel, and pipelined execution architecture, significantly accelerating every stage of large-scale data analysis.

Amazon Graviton5 is the latest generation of AWS Graviton processors, designed for high-performance cloud computing, database workloads, and data-intensive applications, with enhanced hardware execution capability. AWS EC2 M9g instances powered by Graviton5 deliver up to 25% higher compute performance than Graviton4. For database workloads, performance gains reach up to 30%, providing a powerful foundation for real-time analytics, complex SQL processing, and high-concurrency data services.

Amazon Graviton5's performance gains are driven by coordinated upgrades across multiple hardware dimensions. Each chip features up to 192 CPU cores, providing significantly greater high-density parallel processing capacity. Compared to Graviton4, CPU clock speed rose from 2.8 GHz to 3.3 GHz; total L3 cache capacity expanded to 5 times that of the previous generation, while per-core L3 cache increased to 2.6 times; inter-core communication latency was reduced by up to 33%, significantly improving data-exchange efficiency in multi-core parallel workloads.

On the data-transport front, Graviton5 also introduces dedicated enhancements for distributed workloads. M9g instances see up to 15% higher average network bandwidth and up to 20% higher average AWS EBS bandwidth; the highest-end instances deliver twice the network throughput of their previous generations, substantially alleviating I/O and network bottlenecks in large-scale data-processing scenarios.

Building on these hardware advantages, Apache Doris has optimized key execution-path operations, including scanning, joins, aggregation, sorting, and data exchange, enabling it to efficiently support high-throughput queries, cross-node data shuffling, remote data access, and complex analytical workloads in the cloud. Benchmark results show that, under the same Doris version and data scale, Apache Doris running on Amazon Graviton5 significantly outperforms its x86 counterpart, delivering up to 78% higher query performance in standard test scenarios.

Leveraging the robust hardware capabilities of Amazon Graviton5, Apache Doris gains significant performance advantages in parallel processing, cache efficiency, and multi-task coordination:

• Multi-Core Parallelism Aligned with the MPP Execution Model: Apache Doris breaks down complex SQL into multiple parallel task streams and leverages multi-core architecture to accelerate data scanning, joins, aggregation, sorting, and related operations. With more CPU cores and greater parallel throughput, Amazon Graviton5 provides the hardware foundation Doris needs to support high-concurrency workloads, complex queries, and continued performance scaling.
• Large Cache Capacity to Accelerate Hot Data Access: Analytical workloads such as hash joins, aggregations, and sort-merge operations frequently access the same datasets. Amazon Graviton5's expanded cache capacity enables rapid retrieval of hot data and cuts wait cycles, helping Doris maintain stable, efficient performance when executing complex SQL queries.
• Reduced Inter-Core Latency for Better Task Coordination: Apache Doris uses a pipelined execution model that relies on efficient coordination across local computation, data redistribution, network transfer, and result aggregation. By reducing inter-core communication latency, Amazon Graviton5 lowers the overhead associated with multi-threaded, cross-core data exchange, significantly boosting the overall speed of high-concurrency queries.
• ARM Vectorization to Boost Per-Core Compute: Apache Doris has continued to strengthen support for ARM architecture, introducing vectorization optimizations for the NEON and SVE instruction sets. Across columnar scanning, predicate filtering, expression evaluation, string processing, and aggregation, vectorized execution dramatically increases per-unit-time data throughput and maximizes the performance potential of Amazon Graviton5.
• End-to-End Data Pathways Powering Distributed Workloads: The efficiency of distributed data analytics depends on a combination of capabilities, including compute throughput, storage I/O, network bandwidth, and inter-node data exchange. Amazon Graviton5 brings comprehensive upgrades to memory, I/O, and networking, ensuring stable operation across all Doris distributed workloads, from distributed computation and data shuffling to cross-node reads and multi-cluster analytics.

Beyond these performance gains, VeloDB has also achieved strong business growth through AWS Marketplace. As of year-to-date 2026, VeloDB's gross subscription sales (GSS) on AWS Marketplace in just six months have already exceeded its full-year 2025 total by more than 50%, with full-year year-over-year growth projected to triple.

VeloDB's rapid business growth is driven by two main factors: First, AWS Marketplace streamlines the procurement process, enabling customers to consolidate software fees into a single cloud bill. Secondly, the platform ecosystem allows VeloDB to efficiently reach global enterprise customers, while AWS's trusted brand helps shorten evaluation and purchasing cycles.

Earlier this year, VeloDB joined the AWS Marketplace Seller Prime program, which brings multi-dimensional support: product-led growth (PLG) strategy guidance to refine on-demand pricing and free-trial conversion funnels. The program also offers up to $40,000 in Marketing Development Funds (MDF) to precisely target in-ecosystem prospects; and customized go-to-market (GTM) reports plus dedicated customer-acquisition support from the Seller Prime team—all of which work together to continuously optimize storefront presentation, improve trial-to-paid conversion rates, and build a sustainable growth system.

VeloDB's experience demonstrates how cloud-native software providers (ISV) can leverage AWS Marketplace and Seller Prime to accelerate customer acquisition, expand global reach, and build sustainable growth engines.