MetaFluent Publishing Solutions

Many trading firms need to publish internally generated data streams. In fact, some create more internal data than they consume from the markets. Traditional approaches to internal publishing involve numerous hops and components. This increases latency, complicates configuration, and provides too little control over data flows.

MetaFluent offers LAN-based publishing solutions that are simple and low-latency, yet functionally rich.

Publishing interfaces

There are currently two ways to publish through a MetaFluent solution: from Excel or via an API.

The vendor-neutral MetaFluent Add-In for Excel makes it easy for the designer to publish Excel-generated data into a variety of external systems. Data publishing scope can be limited to Excel, for testing purposes, or broadened to external systems. A simple and intuitive VBA API makes development a snap (support for .NET in development).

The [LINK] MetaFluent JMS API provides a simple, vendor-neutral paradigm for publishing data streams in Java, .NET, or (soon) C/C++ applications. JMS MapMessage provides a familiar and accepted name-value pair format for data.

Topologies

These interfaces connect to a MetaFluent server cluster, which provides several options for routing user publications:

Local workgroup publishing. Publishers and subscribers connect to the same MetaFluent server, providing very low-latency distribution.

Routed publishing. User publications are quickly routed to destinations such as RMDS (SSL Inserts to a P2PS, managed publishing to an RMDS Source Distributor) or other middleware.

Server-side configurations include MetaFluent for RMDS and the upcoming [LINK] MetaFluent Managed Content Hub.

Caching models

The topologies above can take advantage of two different caching models. Both models can be used simultaneously by the same MetaFluent server instance, with the caching model determined by topic.

Unmanaged cache - In this mode, users can publish whatever they want, creating data models on the fly. While this approach lacks administrative control and data-model validation, it is popular with some trading desks that place a premium on flexibility.

Managed cache - In this mode, administrators control the data models used by publishers. This control ensures that fields have the same meaning across a business group or enterprise and prevents common problems such as users publishing the wrong data types in fields. The MetaFluent client library or MetaFluent Add-In for Excel validates that messages conform to the data model before they are published. A managed cache also enables logic to be applied to the data in order to transform or enrich it. For example, interal rates from different groups can easily be combined to create real-time yield curves.