What is ArcticDB?

ArcticDB is a high-performance, serverless DataFrame database purpose-built for quantitative data science, time series analytics, and large-scale numerical workloads. It is designed to process billions of rows and hundreds of thousands of columns in seconds, with a Python-native API.

Data goes in as a Pandas DataFrame and comes out as a Pandas, Polars or PyArrow DataFrames - no data transformation required.

Does ArcticDB provide data?

No. ArcticDB is a database, not a data provider. It stores, versions, and serves data that you ingest from your own sources. What data you store and where you source it from is entirely your responsibility.

How does ArcticDB differ from other time series databases?

Unlike traditional time series databases such as InfluxDB, TimescaleDB, Databricks, QuestDB, and kdb+, ArcticDB is a DataFrame database - meaning it is designed around the Pandas DataFrame as its native data model.

Key architectural differentiators include:

• Serverless architecture: ArcticDB requires no dedicated database server. The client reads and writes directly to object storage, eliminating server bottlenecks, single points of failure, and the need for ongoing server maintenance or patching.

• Python-native, DataFrame-centric design: ArcticDB provides an intuitive Python API backed by a high-performance C++ data processing engine. It is DataFrame-in, DataFrame-out by design - no intermediate query language or schema definition is needed.

• Bitemporal versioning: ArcticDB versions data automatically on every write, enabling point-in-time queries, reproducible research, efficient daily appends, and historical corrections. This makes it particularly well suited for research, regulatory tasks, and backtesting.

What are the typical use cases for ArcticDB?

Quantitative research and data platforms

ArcticDB was built by quants, for quants, born out of the real-world data challenges faced by quantitative researchers at Man Group, one of the world's largest systematic hedge funds.

• DataFrame-native API

  Researchers think in DataFrames, and ArcticDB speaks Pandas natively, so there is no impedance mismatch between analytical workflows and the storage layer.

• Automatic versioning

  Every write creates a new version, enabling full reproducibility of any prior dataset state.

• Serverless architecture

  No database servers to manage and no DBA tickets. ArcticDB sits on top of object storage or LMDB, scaling horizontally without operational overhead.

• Fast time series query performance

  Purpose-built for the column-oriented, time-indexed queries that dominate quantitative workflows.

• Seamless scalability

  Handles everything from small research teams to petabyte-scale production datasets without architectural changes or re-platforming.

• Cross-team interoperability

  Provides a shared, consistent, versioned data layer so that research, risk, and execution teams all read from the same source of truth, eliminating data silos and reconciliation overhead.

Self-service analytics and data democratisation

ArcticDB acts as a central data repository that democratises data access across an organisation. Its simple, Python-native interface to large-scale datasets enables users to independently query data, generate insights, and produce reports without requiring specialised database expertise. Note that you can only do this using cloud storage, for example S3.

Systematic trading infrastructure

ArcticDB supports end-to-end systematic trading workflows, from data ingestion (including tick data) through risk management, analysis, and live trading - providing consistent, low-latency access to data across all stages of the pipeline. ArcticDB supports nano second granularity, but is optimised for batch workloads.

Who uses ArcticDB?

ArcticDB's primary user base consists of quant researchers, quant engineers, data engineers, data scientists, and analysts working with large-scale financial and numerical datasets. Its intuitive Python-native design also makes it accessible to risk management, portfolio management, and other data-driven teams who need efficient access to analytical data. ArcticDB scales from individual users to enterprise-wide data platforms.

Can ArcticDB handle large-scale quantitative financial data?

Yes. ArcticDB was built at Man Group specifically to handle the scale and complexity of front-office quantitative research. It scales from small teams to petabytes of data and can process over a billion rows per second.

What storage backends does ArcticDB support?

ArcticDB is verified to work with:

• Google cloud storage

• AWS S3

• Azure Blob Storage

• LMDB (local storage)

• In-memory (for testing and experimentation)

• MinIO (Linux)

• Pure FlashBlade S3

• Scality S3

• VAST Data S3

ArcticDB supports any storage backend that implements the S3 protocol, meaning many S3-compatible object stores should work out of the box. Support for additional storage backends are planned.

Is ArcticDB open source?

ArcticDB's source code is publicly available on GitHub, but it is licensed under a Business Source License (BSL). It is free for personal, academic, and non-commercial use. A commercial licence agreement is required for any business use - including research environments, development, or any activity from which economic benefit is derived.

How does ArcticDB handle data security and access control?

ArcticDB is a client-side library that delegates authentication and access control entirely to the underlying storage backend. When using S3, Azure Blob Storage, or other supported backends, data security is governed by the permissions, encryption policies, and access controls configured at the storage layer. ArcticDB does not implement its own authentication or authorisation layer, meaning security is managed through your existing cloud infrastructure and governance frameworks. Fine-grained permissioning is a long-term roadmap item.