Arenadata One (AD.One)

The main articles are:

• What is IaaS
• PaaS - Platform As A Service

2025: Arenadata One Cloud Platform Announcement

On April 28, 2025, Arenadata introduced Arenadata One (AD.One), a cloud-native platform for data processing and storage. It is designed to work in an elastic cloud infrastructure under the control of the Kubernetes orchestrator and is compatible with object storage based on the S3 protocol. The platform provides ease of scaling, flexibility of management, and versatility with data in a variety of formats, combining both analytical and transactional workloads.

𝓲

Фото: Arenadata

According to research by Arenadata, in large companies, analytical DBMSs are idle 40-50% of working time. This is due to the fact that classical technologies for analytical and transactional problems, placed on bare-metal for maximum performance with homogeneous planned loads, require the reservation of equipment in the cluster at floating peaks, which in turn leads to hardware downtime and an increase in the cost of storage ownership.

According to the company, Arenadata One combines a multi-tenant architecture with a low T2M (time-to-market) DBaaS solution. The platform provides isolated service to various groups of users within a single infrastructure, while offering technologies as part of an on-demand database deployment. In the DBaaS model, the cloud provider takes over the management of infrastructure, configuration, scaling and database maintenance, providing users with convenient access to the service through the API and web interface. It is noteworthy that DBaaS functionality is available not only in the public cloud, it can be deployed in the customer's private infrastructure. This provides additional benefits when building a data platform.

𝓲

Фото: Arenadata

{{quote 'author
= noted Alexander Vinogradov, director of the Arenadata' Arenadata One cloud products business area - this is a different approach to storage infrastructure in the data economy with the widespread introduction of artificial intelligence and machine learning technologies. Arenadata One has affordable auto-scaling scenarios, the ability to reuse computing power, store in a single unified storage with support for the S3 protocol, as well as payment models for the used resources of a cloud provider (pay-as-you-go) and allotment of costs between structural units of the company that use data. The introduction of modern technologies into the business, such as large language models (LLM), machine learning operations, the industrial Internet of Things, as well as continuous increases in the cost of data ownership require new approaches to building a storage infrastructure. The hybrid architecture of the Arenadata One platform gives customers the freedom to choose whether to deploy everything in the private cloud, use the provider's cloud, or find the best balance between different types of placement for their tasks.}}

𝓲

Фото: Arenadata

AD.One solves problems such as:

• building Lakehouse;
• real-time analytics;
• Storing vectors and features
• On-demand database.

Arenadata One features include:

• Architectural separation of Compute and Storage. Flexible and efficient approach to hardware utilization through isolation/resource sharing and Compute auto-scaling, including scale-to-zero.
• Integrated cloud-native OLTP and AI storage. Scale-out components that use RAM and Disk for real-time analytics, vector storage, and features.
• Decline OpEx and. CapEx Reduce storage and processing costs, increase ROI (), ROI lower total cost of ownership (), TCO and reduce time to market (T2M).
• Multi-tier architecture. A mechanism for thermal storage and processing of data based on user priorities and business requirements.
• Data Governance. An integrated management tool for model, quality and data connectivity that is inextricably associated with Compute and Storage.
• Storage S3. A single unified storage standard that allows you to store data and backups in the same Storage with the ability to analyze, including over backups.
• Hybrid Cloud. A service for building hybrid cloud storage, where part of Compute and Storage is located in the customer's infrastructure, and part is located at the cloud provider.

The Arenadata One data platform combines two complementary versions into one self-service management console: Arenadata One OLAP (Online Analytical Processing) Edition and Arenadata One OLTP (Online Transaction Processing) Edition.

Arenadata One OLAP Edition supports a wide range of analytical tasks:

• Interactive data analysis and regular reporting.
• The full cycle of Data Science is from experiments to industrial ML.
• Federated requests to distributed sources.
• Predictive modeling and risk analytics.
• Building complex financial models.
• Processing high-load analytical workloads.

Features of the OLAP version:

• Optimized work with analytics in a single environment due to the connectivity of all platform components with a minimum of manual settings.
• Fast start-up and scalability-Automate with Kuberneter Operator and manage the entire system from a single console.
• Increased security - implemented authorization for Apache Ranger and advanced monitoring.
• Seamless integration of all system components configured for out-of-the-box collaboration.

Computing cores:

• Apache Spark is a universal engine for data processing, machine learning and ETL;
• Apache Impala is a high-performance MPP engine for Big Data SQL analytics.

Integration tools:

• Trino - federated requests to heterogeneous sources;
• Apache Airflow - orchestration of complex data-pipeline;
• Apache Kafka - real-time event streaming;
• Apache Flink is a complex stateful transformation of streaming data.

Each technology is tailored to meet specific challenges within a single platform, providing optimal performance, architecture flexibility, ecosystem openness, and solution scalability, and is open source software.

The Arenadata One OLTP Edition is based on Neon and Picodata technologies. Neon is a serverless implementation of PostgreSQL, created to work in elastic cloud environments as a fully managed service. Retains full compatibility with the "vanilla" version of PostgreSQL.

Among the functionality of Neon are dynamic auto-scaling of resources, the scale-to-zero function for saving resources, branching of data, in particular for optimizing CI/CD processes, centralized management of Postgres instances through a single UI interface, support for Postgres Extensions for working with vector data and features.

Picodata is a distributed computing system in RAM that helps implement real-time data processing scenarios, including hot storage of vectors and features for working with large language models (LLM), as well as tokenization of sensitive data, including for compliance with requirements No. 152-FZ when working with personal data in hybrid architectures.

It also plans to implement its own Multitiering Data Migration technology for data migration between systems with different performance and storage temperatures, which allows you to warm up and provide quick access to data.

Use Cases:

• Data analysis: real-time analytics, data marts, behavioral analytics, telemetry collection and analysis, on-demand database.
• Storage of vectors and features for AI: Copilot's, AI&RAG chatbots, semantic and hybrid search, etc.

The platform has integrated monitoring, orchestration, and data management tools to implement the Data Governance concept. Increased attention is paid to information security when working with data - the role model allows you to configure access rights at the levels of lines and columns, there are built-in authorization and authentication tools, integration with popular SIEM systems.

Arenadata One is fully ready for a commercial start.