The Scaled Risk Platform provides a suite of capabilities for Financial Institutions to easily integrate, secure, and analyse Data.
DYNAMIC AND Flexible Data Schemas
We have conciliated Big Data and traditional software. With Scaled Risk, your organization is able to keep on using traditional IT modeling while taking advantage of a powerful Big Data layer. Prior data models can be reused for a smooth transition to the Big Data era. We offer a wide range of formats to retrieve models, from database schemas to Java classes and of course our internal format.
Our data model is fully customizable and embeds a constraint engine. It is possible to define relations between entities, keys, types and even to use hierarchy. ‘Unstructured’ data can be imported into Scaled Risk without any prior type definition.
Read more and discover how Scaled Risk leverages a flexible data metamodel
Scaled Risk gives you direct access (restricted by rights) to the whole data set of your organization : market data, trades, positions, reports, contractual document, Excel Sheet… So you are able to query these documents with a simple and humanly readable language, just as you do when using a web search engine.
Queries can be refined and results sorted or filtered with simple from the GUI and exported to your favorite Spreadsheet tool (i.e. Excel).
We offer a powerful integrated indexing capability and a human readable query engine so any Financial Data can be found within seconds among billions of record and various documents.
- It is not necessary to know about the underlying model to query it.
- The exact result or set of results can be found at once if the query is accurate enough or reached step by step by refining the query and mining the data.
- The mining is done directly on the whole dataset, no smaller dataset are created as intermediate steps, so the process is Big Data end to end
The queries are fast and human readable, so IT skills are not required to extract data. If this task is still delegated to an IT team, queries can be executed very efficiently. The IT team can dedicate its time to more productive, instead of spending up to 40% for extracting information on Front-Office and other departments’ requests.
In Scaled Risk, everything is streamed in Real-time and processed, including indexation. The indexation mechanism does not require a separate infrastructure; no bandwidth is consumed to send the data to some other location.
Read more about what makes the search feature of Scaled Risk unique
AS-OF-DATE & VERSIONING
Every modification is kept and time-stamped, Scaled Risk does not allow deletion. You now have a unique system that gives you at the same time, capabilities to comply with regulation constraints and to access an unlimited data history for your business.
Build a time-stamped golden copy of your data, at your own pace. Data Quality can not be a once-off cleansing; it needs constant monitoring and improvement. We propose a new way of managing “Bad Data“. Because they are as valuable as any data, Scaled Risk does not reject them and instead mark them so they can be fixed through a comprehensive workflow.
Real-Time & Distributed In-MEmOry OLAP
Running a real-time query can be done from the Graphical User Interface, by simply using the same queries as above. Results can be rendered in various objects such as ticker, tables and charts (e.g. time-series curves or pie charts).
It is not much more difficult to define a Real-time P&L aggregation from a stream of incoming trades that are generated by external and heterogeneous systems.
Our OLAP cube offers best performance in real-time, based on in memory distributed aggregation capabilities. Latency can be enhanced by adding extra nodes to your Hadoop clusters (horizontal scalability), data distribution is built-in with extra administration tasks needed. Our OLAP cube provides full choice for dimensions, multiple aggregations per cell, best performance for initialization and low latency for real-time updates.
Read more about the Scaled Risk scalable and distributed In-Memory Analytics engine
Scaled Risk embeds an event-driven framework that enables real-time capabilities as defined for financial applications.
Scaled Risk approach is to provide best performance and lower latency on nominal case. In case of resource shortage Scaled Risk sacrifices latency, but optimize throughput and performance. Scaled Risk always guarantees reliability and completeness of all messages.
Scaled Risk automatically adapts its real-time transport mode, depending on resource availability.
Read more about Scaled Risk real-time architecture
Scaled Risk guarantees the integrity of transactions thanks to the world acclaimed HBase. HBase is doubtless the best distributed database of the Hadoop ecosystem regarding reliability, real-time processing, robustness and node recovery time after failure. Thanks to very strong references in the web industry, our involvement and commitment for enhancing it, we have chosen HBase to build our application.
Once data has been imported to Scaled Risk you know that that it will stay forever even in the case of hardware failure. And with Scaled Risk, data are available as soon as published, in real-time.
Anyway you will forget about the underlying technology, and appreciate:
An unlimited data storage capacity for a full data history and audit trail.
Transactions acknowledged by the system even at high input rates (up to several millions per second).
Massive parallel calculation capabilities that enable real-time processing of data streams
Scaled Risk offers a set of machine learning algorithms that can be directly applied to a various range of financial use cases form mood trading, to fraud detection.
Machine learning can be used at the heart of Scaled Risk which means with instant access to data history, applied to incoming real-time events. End-user can provide immediate feedback which is taken into account to adapt and make Scaled Risk smarter every day.
OPEN tO STANDARDS
We have included several standards to Scaled Risk, FpML, known for its time-consuming implementation. We provide the technical stack to handle FpML applications: trade repository or communication with clearing houses. We are also able to integrate your custom or FpML-inspired markup-language.
We are also Bloomberg Open Symbology facilitators.