AB2000 Software is a privately owned Canadian company that develops low-cost DMA technology designed specifically for high-frequency high-volume algorithmic trading. Our low-latency electronic trading platform is a complete trading solution that can be used by companies involved in high-frequency trading arms race. Cost of the platform is significantly lower compare to products with similar functionality offered by other technology providers (like FlexTrade, ORC Software, Progress Apama just to mention a few).
The trading platform features functionality that allows a broker
that sponsors DMA to have full risk control over the order-flow without
introducing extra delays to execution. This makes the solution fully compliant with new rules for DMA trading adopted in the USA (similar
rule is coming to the Canadian market soon).
The platform includes a low-latency FantasTickTM Ticker Plant
to process direct market data feeds , high-performance FIX Engines
with drop-off capabilities, Algorithms Hosting Engines
, a centralized Risk Management Server
and a Trading Controlling Console
to monitor and manage trading activity and to supervise state of the entire trading system. The platform is a distributed scalable solution that can run on a single PC or on a group of PCs for the purpose of load distribution.
During trading the platform does not deploy a traditional database. That is why it is extremely efficient and is not affected by problems typical for database-oriented trading systems. Significant spikes in market activity can be quite challenging for any traditional relational database when order flow in a trading system reaches several thousand a second.
The platform also features market simulation environment where simulated orders execution is driven by feed replay using recorded raw market data feed files. This functionality can be very handy for strategies back-testing.
One very useful feature of the platform is a parameterized managenemnt of a queue of open ordes . By specifying settings that define open orders price
levels and size distribution it eliminates the needs to track open orders in a
strategy code or to manage them explicitly using methods like “send
order” or “cancel order”. Relieving a developer from necessity of orders
management in the code makes a strategy development much simpler, considering that management of layers of open orders usually takes a lot
of efforts from coding perspective.
Writing trading algorithms for hosting by the platform is quite simple and requires only basic knowledge of C++
. The system is designed in the way that allows a trader/developer to concentrate on a trading business logic instead of fighting technological and exchange-specific issues. The platform API is exposed as a thread-safe event-driven callback interface that can be invoked from multiple threads. A trading logic is implemented inside strategy DLL
modules hosted by the Algorithmic Hosting Engine
process. The hosting engine deploys high-performance threading model that provides superior efficiency and allows sending of several thousand orders per second
to multiple destinations with low CPU usage. Direct market data feed handlers can be also hosted inside the trading engine process to avoid latency associated with market data feed dissemination. The Algorithmic Hosting Engine can host hundreds independent strategy DLLs at the same time.
The Trading Controlling Console
exposes a very ergonomic GUI
interface that with a click of a button allows a trader to control
hundreds of trading strategies running on multiple servers and to
supervise the health of the underlying infrastructure.
You can download technical documentation and a demo project from this FTP site
using "guest" as user ID and password.
We offer free no-obligation two-month trial and help with the platform initial deployment. We believe you will be impressed by efficiency of the platform and simplicity of its deployment.
For more information contact Alex Bevziouk
We also offer hosting of servers at exchange co-locations in the USA and Canada what allows our clienst to access liquidity with sub-millisecond latency.