Flexible scalable framework
Our framework is based around a generic central event processor (CEP). From this core any number of different business functions can be attached in the form of modules. This gives the client the ability to process all the business requirements of an event, for example a trade, within a single integrated application.
Utilising this model reduces the data processing errors that can occur when running separate business applications. The modular-based system allows the user to customise the application by choosing the modules that apply to their business area. Flexibility and adaptability, by easily adding or removing modules in the future, are key features of the application enabling it to respond to your evolving business.
Flexibility to customise
The framework has been designed to fit into the business, not the business fitting into the framework. If today you are trading futures, and tomorrow you want credit default swaps then our framework can do that without the need for you to adjust the core infrastructure, the pricing model, the order routing. That is all part of the flexibility in the framework design.
Data exchange between business functions
Communicating within business departments is important. When a trade is executed the middle and back office need to know about this and it should not be a delayed feed. Real time is crucial so that the exposure can be calculated and Profit And Loss can be managed. In the Celer framework, each module can communicate instantaneously with any other module via simple exposed internal Application Programming Interfaces (APIs).
Upgrades to individual modules without disruption
Individual internal Application Programming Interfaces (APIs) have been applied to each module allowing them to operate independently, whilst still having the capability to asynchronously exchange data between each other via the CEP core. Having module-specific APIs enables updates to a module’s internal business logic without affecting the other modules that it communicates with. This allows each module to run on a different release cycle.
Scalability through clustering
The application can be installed over multiple servers providing a clustered network of frameworks that communicate with each other. This gives the Celer application scalability as well as enabling all event data to be distributed seamlessly between servers, regardless of geographical location.
Next generation database enabling high speed and large datasets
The Celer framework is able to support multiple types of databases. This allows the framework to be adapted to the needs of your application. We have done extensive research into various types of database solutions. As a result, we have a preferred method of using an alternative non-Relational Database Management System (RDBMS). This next generation database (Big Data) provides the framework with better performance and scalability.
Downtime prevention within the Celer framework is achieved through replication and re-direction of data. This is made possible by the independence of the APIs in each module as well as the clustering within the application. If a server in a cluster fails, the data traffic is automatically re-directed to another server within the same cluster and the service resumes without interruption. Further, the Profit And Loss of application downtime due to the loss of the database store is reduced by the replication of the data over multiple nodes. The independent operation of modules attached to the central CEP engine and clustering allows the administrator to correct a failed module or node without interfering with the operational functionality of the entire framework.
Ability to correct failed modules whilst the application is still operating
With our distributed installation you can keep separate failed modules and still keep you system up and running. So in the event of a pricing engine going down to one exchange you are not out of the market for Profit And Loss and clearing. The problem is isolated and can be fixed whilst the system is up and running.