What is Carbon09?

Carbon09 is a set of financial markets primitives. These primitives enables developers to decompose and build financial analysis applications for the cloud . Primitive components create immediate value for developers by enabling them to design applications for speed, scale, and fault tolerance. Consider the carbon lattice materials of today: light, incredibly strong, yet flexible like no other material. Carbon09’s primitives offer the future of financial markets software today.

For all the media coverage of the vast sums of money in the financial markets, the technology remains fairly primitive. The limits of yesterday’s technology result in systems that remain static after construction. Yesterday’s software construction resulted in systems cumbersome to build and difficult to adapt to the new post-Dodd Frank financial markets landscape.

The brittle components from which existing systems are built constrained the imagination of quantitative analysts, strategists, financial solutions engineers, and technical architects. From time to time the media reports symptoms of the stressed and outmoded building blocks of yesterday’s design.

 

What is the Carbon09 QuantLib distribution?

Among other marvels we’ve ported the ~955,000 lines of QuantLib C++ code to Asm.js. As a result we can run with no garbage collection and very low variance code execution. Asm.js executing as an AWS Lambda function enables very high performance at scale compared to non-native code while enabling simple integration with real-time data events.

Blazing Fast C++, but also Asm.js, and AWS Lambda functions

Why Not? Our research found many attempts to hard-code C++ libraries like QuantLib to Java, C#.

How does Carbon09 augment the performance and Utility of a FOSS Quantitative Finance libaries?

As FOSS QuantLib source code can be adapted to various use cases. The brilliant minds contributing to the project (the World!) create and evolve the library and a rapid pace. Few enterprises could build such a library. And none can adapt as quickly to changes in the industry.

As an essential input to of the Carbon09 Marketplace, the fact of Quantlib.org’s  business friendly MIT license enables Carbon09 to stand on the shoulders of giants, but also to do some genetic engineering.

High Performance Scheduling and Management of Spot Instance Powered Clusters

Quantitative finance develops new a new kind of DevOps–a Devops built around the unique needs of structured finance. Carbon09 enables developers to execute with low fraction, low cost, and high performance.

Carbon09 C++ QuantLib Distribution

For the ultimate in high performance execution we’ve built QuantLib to leverage high performance computing on AWS. Our distribution is built with high performance Community Edition of Intel MKL libraries.

Carbon09 Optimized for AWS GPU 2 Instance Type

g2_gpu_model_4Rigorously built and tested we compiled the Carbon09 library for parallel execution on the many cores of the Nvidia chips.

AWS C4 Instance

We’ve optimized for native Intel chip capabilities such as Intel AVX, AVX 2, predictive branching, caches, and more.

AWS I2, D2, and other Instance Types

Based on your particular application and workload we analyze the characteristics of your workload as it executes in real-time, and our service compiles, builds and deploys your native code optimized for your workload. We analyze the performance data and can help you to continue to optimize for performance and cost,

 

Value Proposition
Developers begin with a Carbon09 distribution of QuantLib optimized for specific AWS instance types and pre-installed on the AMI.

 

Developers can simply launch, link, and run. We’ve automated the build process and developers can leverage our simple build and Deployment process to schedule and run a fleet of spot instances.

 

AWS C++ Integration for Fintech

We’ve augmented Carbon09 examples to include how to leverage AWS using native code. Working examples illustrate how to leverage the blazing speed of native code to read data from Kinesis, Redis, SQS, and S3.