eXtremeDB for Financial Systems
McObject offers a reference application showing how in-memory, high availability data management can be leveraged to implement high throughput algorithmic trading. The reference package includes the database design, sample data, and easy-to-use scripts to see the applications in action. A full-featured, distributed database system underlies the system, guaranteeing data integrity (through transactions) and openness (SQL, ODBC, C/C++ interfaces). Source code for the reference applications is available when you request an evaluation copy of eXtremeDB.
Hedge funds, pension funds, mutual funds, and other institutional investors are invited to use McObject’s platform as a quick-start kit for developing high volume automated trading with unsurpassed performance.
Advantages of McObject’s Algorithmic Trading Reference Application
- Increase performance many times over through elimination of RDBMS-based bottlenecks
- Leverage 64-bit OS to grow business volume or choose 32-bit architecture for cost-effective, smaller-scale trading
- Software is optimized for the latest multi-core hardware platforms
- Lock in fault-tolerance using database replication with automatic failover
- Enhance modularity, distribute processing intelligently and reduce network traffic overhead via decentralized design based on distributed local caches
- Openness and data integrity guaranteed through McObject’s underlying database system (supports ACID transactions, standard SQL and native C/C++ programming APIs)
- Available for 64-bit Linux or 64-bit Windows (Read the Technical Specifications)
- “Master” in-memory eXtremeDB database receives incoming data stream to obtain basket of relevant securities
- DBMS High Availability (HA) mechanism distributes data stream to replica nodes, where replica databases provide distributed local cache instances
- Trading stations query these local caches (replica in-memory databases) exclusively to identify securities to buy/sell/sell short, etc.
- Traders have no interaction with Master database, which significantly reduces “round-trip” network traffic and frees Master node to manage, sort and distribute the incoming data stream.
- Number of replica nodes is configurable to match the number of traders (real or “black box”)
- In event of Master node hardware or software failure, a replica node can assume the Master role to allow trading to continue uninterrupted