Fractal Executive Summary

Fractal delivers quantum computing results using current hardware. Operating at 200 million transactions per second and virtually eliminating I/O bottlenecks, Fractal enables legacy applications to be replaced at one-tenth of their expected cost.

New applications, particularly in AI, can be built and deployed on minimal hardware within a quarter, operating between 1,000 and a million times faster than conventional technology.

Fractal is the next evolutionary step in microservices and object-oriented programming. Instead of microservices, it delivers micro apps—entirely self-contained technology stacks that achieve outcomes unattainable with current technology.

Fractal is most effective in environments where speed must increase by a factor of 1,000 to 1 million, making it a powerful tool for AI acceleration, machine learning, massive data processing, and real- time ETL.

• Most applications cost about half the cost of the legacy version of the application.
• Most applications are delivered into production in a single business quarter.
• Applications generally run from 1,000 to 1 million times faster than the legacy version of the application.
• Applications do not require large data centers; they can run on hardware you can hold in your hand.
• Most applications use 80% – 90% less storage than the traditional app they replace.
• There is no need for Oracle, VMware or other legacy technology.
• Application development and deployment will only require the skills of a few Node.js full stack developers, not an army of application developers and DevOps teams.
• Applications will run on any hardware with a Unix instance, in the cloud, in a data center or embedded device with no code change.
• Applications typically reduce the cost of cloud hosting by 50% – 90% because there is no need for expensive cloud value added services.
• Apps are fully distributed with required security, meeting current requirements.

Fractal is not new. Its basic concepts were developed 35 years ago in a United States government-funded research project. Early versions of the technology required expensive custom hardware. Over the past 35 years, technology has evolved to where a Fractal app now runs on general purpose computers. With the advent of Intel NUCs, and similar inexpensive compute platforms, Fractal enables significant savings for enterprise application development and deployment.

Fractal apps have been developed and deployed in large electric and water utility billing and customer care systems for over 7 years. In 2022 and 2023, Fractal built a custom CRM-like system with 1.7 billion records, operating in real time, on a couple of Intel NUCs – in 90 days.

In 2024, as a demonstration project, Fractal ingested the entire Federal Election Commission contribution database—680 million records—showing that this massive government system, which required an enterprise-level data center and struggled to process 500,000 transactions per week, could run on a handheld computer at quantum speed, with no data center needed.

Fractal concepts enable system integrators and other solution providers to be trained in their specific application and used across any industry.

1. Locality of Reference

   Locality of reference means a micro app creates data pipeline-processing for the computer’s CPU. This enables applications to run at “CPU speed/silicon speed,” which is orders of magnitude faster than I/O-bound system speed which is the traditional limit on application performance.

2. Locality of Logic

   Locality of logic is a concept where many of the processing and calculating steps in any app are pre-built into the software stack, somewhat like a compiled application. This enables massive speed, takes advantage of locality of reference and dramatically reduces code complexity and size.

The optimal manner of evaluating Fractal is to build a parallel app to a current legacy app one wants to either replace, test, or augment with new features too complex or too expensive to add directly to the legacy app.

The Fractal app is typically built in 90 days or less. That Fractal app runs in parallel with the current legacy app for multiple quarters. The parallel Fractal app continually reconciles its results with the legacy system. Every line, every invoice, every transaction is compared to assure 100% fidelity of outcome. Issues in both the parallel app and the legacy system are quickly identified and resolved.

Parallel application development requires limited customer resources. Typically, there is a need for a business user to be available for 3 – 4 hours a week by phone to answer questions. There is also a need for a programmer who understands the data in the legacy app for 3 – 4 hours a week. There is no need to see or touch the legacy source code.

The parallel application is initially developed with a small data set (a Fractal) on the order of 25,000 records so that the app can be quickly developed and tested. Once that Fractal is operational,   millions,  billions  or   more   records   can   be   added.   The  Fractal framework automatically does the heavy lifting of scaling the application to full production scale. Thus, the name Fractal.

The parallel application methodology provides a low risk, low cost path for testing, augmenting, and eventually replacing large expensive mission critical enterprise applications, at a pace that is customer controlled.

Fractal apps run in a truly distributed environment with the same code base whether on a computer, embedded device, at the edge, in the cloud or on a mobile device. Fractal applications deliver quantum speed outcomes in real time on current hardware.