All FeedOS Server Components are based on QuantHouse’s FeedOS Framework. Developed in C++ and natively multi-threaded, the FeedOS Framework was specifically designed for market data processing and dissemination, focussing on high throughput and ultra-low latency.
Below are the main FeedOS Server Components, used in QuantFEED Solutions.
FeedHandlers decode and normalize market data updates received from the exchange or from 3rd party vendor feeds. The normalized market data updates are disseminated using QuantHouse’s proprietary multicast FeedOS protocol.
By default, FeedHandlers are configured to broadcast normalized market data using three multicast channels:
It is possible to mix these channels into a single stream in order to preserver the original sequence of events, as received within the RAW feed.
FeedHandlers manage and normalize the Referential data (also know as Static Data or Dictionary) making it very easy for consuming applications to identify their instruments. Intraday instruments creation are fully automated, and disseminated from the exchange towards the clients applications.
FeedHandlers implement recovery mechanisms provided by the RAW Exchange protocols.
All FeedHandlers can be configured in various ways to:
Subscription Servers (also known as FrontEnd servers) are used to aggregate several feeds coming from FeedOS FeedHandlers. These incoming feeds can be either in multicast or TCP, depending on your setup. Then, Subscription Servers serve subscription requests from client applications. Client applications run these subscription requests using the single FeedOS API.
A large choice of value added features can be configured at the Subscription Server level:
Clients entitlement are managed through logins, Entitlement Ids and Flavors (Ref / L1 / MBL / MBO / Historical / etc.).
Deployed on client’s side to optimize bandwidth usage, when tens (or hundreds!) of client applications need to consume market data, and bandwidth usage between client side and QuantHouse Point Of Presence must be optimized. In that case upstream servers are Subscription Servers in QuantHouse infrastructure, and/or local FeedOS Servers Components (FeedHandlers, Subscription Servers, Publication Servers, etc.) in client premises.
The Order Book Consolidation server allows clients to capture and consolidate liquidity across multiple trading venues. It consolidates Real Time market data across Regulated Markets (RMs), Multilateral Trading Facilities (MTFs), Organised Trading Facilities (OTFs) and System Internalisers (SIs) and delivers them to downstream applications via a single consolidated (or composite) market data feed.
Clients can fine-tune the consolidation behavior (cumulated quantity, latest quantity, biggest quantity, latest market timestamps, etc.). TradingStatus, Opening Price, Closing price and auctions data can be enabled/disabled in the consolidated instruments. Clients can configure the universe of instruments to be consolidated, using industry standard referential data fields. This universe can be modified on-the-fly by the client using the Client Administration tool.
The FeedOS Consolidated Server powers QuantFEED EBBO, the fastest and most flexible European Best Bid and Offer (EEBO). It gives insight into the entire liquidity available across all of Europe’s MTFs and primary exchanges, with synchronized timestamps.
The Order Book Consolidation server can as well help clients to be compliant with the MiFID Best Execution requirements, by allowing them to identify the trading venues that provide, at a given point in time, the Best Price to execute orders.
Request a demo today.
A publication server behave like a passive FeedHandler: using the single FeedOS API or Command Line tools, custom instruments can be created or existing instruments can be enriched (with custom values, back office information, codes from other vendors, etc).
Likewise, quotation data (all price-related info: trades, books, etc) can be generated for all instruments (new instruments or existing instruments).
This server can also behave like a Subscription Server by aggregating other feeds (referential and/or quotation data). Downstream feeds can be broadcasted as a whole (in multicast or TCP) or instruments subscribed individually.
Contact us to review the Publication Server features together.
FeedOS market data updates generated by QuantHouse FeedHandlers are stored in binary compressed files. (FeedOS Binary Historical data). All instruments are present. Using the FeedOS API or a Command Line tool, you will be able to replay the data, and ingest it in your own database or backtesting system.
The TREP Adaptor is a FeedOS server which publishes data flowing from the QuantHouse feeds (QuantFEED, ConsolidatedFEED) to the TREP-RT infrastructure, also known as RMDS. The data is exposed as a custom “interactive” service, readily available to any application built with the RFA or UPA APIs. Instruments codes, Referential data and Quotation data fields mapping is supported. FeedOS TREP Adaptor is natively integrated with DACS.
The WOMBAT Adaptor is a FeedOS Server which publishes data flowing from the QuantHouse feeds to a WOMBAT infrastructure. The adaptor provides facilities to build custom record keys from the Referential tags of an instrument, allowing maximum flexibility. The Adaptor is based on Wombat 2.17.46.