Swim Architecture

Distributed, Stateful, Dynamically Derived

  • State
  • Compute
  • Persistence
  • Messaging

Swim Web Agent

Continuous intelligence at scale is hard. Enter Swim.


Act, not Mirror: Swim Web Agents

Swim Web Agents are web-scale and autonomous, digital agents that aggregate, transform, analyze and act upon dynamic and static data at the speed of change.

Unlike digital twins, which only mirror the state of the real-world entities they represent and don’t relate to each other, Swim Web Agents take action on behalf of their entities and continuously communicate their state via URI to each other and any subscriber at network real-time. They are self-contained, self-training objects that possess individual logic and manage their own data streams and persistence.

Swim Web Agents enable general purpose compute, allowing any kind of business logic, not just a limited set of analytical functions, to be processed with streaming and contextual data.


Analyze Then Store

With traditional cloud analytics, databases are queried to process data, update state and derive insights.

In contrast, Swim’s streaming APIs always listen and know immediately when your streaming data changes. Swim’s stateful architecture already has the context in which the state change happens, remembers the previous state without having to make a database query, and continuously re-computes and re-aggregates all your dynamic and static data in concert.

Storing raw, often only ephemerally useful data for later analysis is just too slow for continuous intelligence and real-time decision making. In contrast, Swim immediately performs automated data reduction and analysis of data in-stream, accelerating the time to insights and action by orders of magnitude.


Persist Locally

Modern streaming, collaborative, AI and edge workloads are extremely latency-sensitive, and continuous intelligence takes that to a new level.

Swim has been pioneering a dynamically derived, distributed architecture to power continuous intelligence at scale, slashing latencies by orders of magnitude. Data is stored wherever it is processed, and persistence is kept off the critical performance path. Swim relies on composable cache coherence for persisting high-frequency updates and is able to perform millions of read-modified writes per second.

Optimized for data locality, Swim leverages available compute from the cloud to the edge, independent of infrastructure constraints.


The Only Software You Need

Swim is the first fully-integrated, database-free platform for building and running continuous intelligence applications and was designed with no 3rd party dependencies in mind. It provides messaging without a broker, tasks without a job manager, business logic without an app server and persistence without a database.

Swim provides standardized artifacts, (SDK, Docker), a client and a UI framework that simplify deployments and allow developers to build new continuous intelligence applications or embed streaming UI components into existing applications.

Building upon statefulness and cache coherence as the foundation for unprecedented performance, scale and efficiency, Swim is uniquely architected to power modern, continuous intelligence at scale.

Find out more about how continuous intelligence applications work.