mimicry is a novel method for understanding and reengineering aging software systems, enabled by recent advances in artificial intelligence and by a unique combination of machine-based and human analyses. mimicry extracts business logic such as workflows, dataflows, functions and business rules, and data types and data models from your application. It develops a mental model of the business functions it performs, ready to be explored by business and technical experts.
The model serves as an executable specification, enabling to run selected or all business logic on a cloud-native architecture mimicking the behavior of the aging application. Enhancements to the business logic can be prototyped with agile methods.
mimicry enables enterprises to unchain the hidden assets in their aging software systems, replicating their business functionality in a simplified application using today’s architectures and infrastructures. It empowers companies to innovate faster, at lower risk, and at lower cost.
mimicry migrated selected business logic of a payments processing platform onto a Hyperledger Fabric distributed ledger, adapting the business logic to store transactional data in the distributed ledger. The generated Go program featured an actor programming style and worked as an application distributed over multiple processing nodes.
mimicry analyzed a mainframe batch program, part of a 40-year old transaction processing platform. The analysis flagged a number of data consistency issue within the application environment. The extracted business logic illustrated the feasibility of mimicking logic embedded in these types of applications.
mimicry learned with minimal human assistance from an order book data feed how to manage an order book, and how to create snapshots. The resulting modern C++ program was praised for its clean structure and high efficiency in a review with software architects.
mimicry learned how to decode system execution traces from IBM's z/OS Generalized Trace Facility, enabling automated analysis of resource consumption and performance issues.
mimicry analyzed the business logic of a rule-based system with more than 4'000 business rules, with the processed input and output data having more than a hundred input and output fields. It extracted complex rule sets, and provided advanced tools to analyze, simplify, and cross-validate rule sets.
mimicry extraced the business logic of multiple server-side rendering web applications using technologies such as PHP and Python/Django, and replicated the logic in modern TypeScript/JavaScript for mixed client-side and server-side use. It enabled the integration of the different solutions into one uniform new web application featuring current web standards, support of modern continuous integration/continuous deployment (CI/CD) environments for agile change, and deployment on content-delivery networks (CDN) for improved user response times.
mimicry client engagements surfaced the need to support the design and maintenance of master data as rigorously as the design and maintenance of software code. Using assets from the mimicry tool set, mimicry developed a product supporting the design and maintenance of complex interconnected master data sets with unlimited time-dependent versioning, full audit logging, and configurable data type, quality, and consistency checking.
See mimicry Master Data Designer for details.
Our mission at mimicry is to make business experts and software engineers focus on novel capabilities for the business with new technology, freeing them from disentangling complexity piled up over years. We thereby empower corporations – and entire industries – to innovate and capture new business opportunities.
Dr. Marc Brandis, Founder, mimicry AG
Marc Brandis has been advising enterprises on strategic IT topics for more than twenty years. He teaches at ETH Zurich and serves on the Board of Directors of ETH Juniors, promoting knowledge transfer between academia and industry.