Competencies

The competencies Quanopt personnel cover a wide range of technical skills supported by solid theoretical background. Our main competencies are applied across application areas from embedded system development to IT infrastructure analysis.

Modeling

We have years of experience in creating models which describe requirements, structure and behavior of complex systems. What makes Quanopt unique is the combined application of lightweight modeling techniques, domain specific modeling and formal mathematical methods in a way that enables business users to concentrate on problem formulation.

• In a storage plant optimization project, we used models to describe manufacturing processes, inventory, orders, physical assets, timing and costs parameters and interpretation of precise mathematical calculations for the end user.
• In an analysis project for the Hungarian branch of a multinational financial institution, we used models to describe back-office processes of a bank, resources, IT infrastructure and timing and correctness requirements.
• In a business process diagnosis project, business logic, organizational data, resources and services were modeled with compliance and SLA-like requirements which made the business user capable to define executable rule-based checks.

Data analysis

The application of data analysis in our consulting services is one of our core competencies. The emerging discipline of "data science" is significantly reshaping the fields we are active in. Some examples from the professional experience of our team:

• IT infrastructure and cloud deployment optimization can draw on fine grained historical monitoring observations in a multitude of ways, ranging from resource-level capacity planning to error propagation chain identification and analysis using rare event characterization techniques.
• Business process assurance can rely on known process execution characteristics for risk-conscious process optimization and re-engineering.
• In process modeling and optimization, not only process characteristics, but also the external "load" of the system (e.g. the orders in manufacturing) have to be understood to the level where trustworthy predictions become possible.

For all of these cases, the application of statistical modeling methods is only one facet of unlocking the potential that is present in available data. When targeting real-life problems, exploratory data analysis techniques and increasingly, Big Data techniques are of crucial significance, too. Our company is in the unique position that we can employ data analysis as a core capability that seamlessly blends with and supports our modeling, optimization and domain specific expertise.

Optimization

Resource allocation and scheduling is a problem that needs not only mathematical methods to solve formal equations but multi-aspect modeling of constraints and user requirements to cover performance, robustness and security issues besides business effectivity.

• In complex embedded systems (like a network connecting control units of a car) multiple criteria prescribe how system communication has to be performed to guarantee safe operation.
• In a plant producing low-volume but very high value products with hundreds of configurable parameters, requirements of production and business are often contradictory. The plant has comply with rapidly changing business requirements and customer preferences, while testing and quality requirements have absolute priority to guarantee customer satisfaction.

Infrastructure design

Services with demanding performance and availability requirements have to be deployed on infrastructures that are properly sized and are equipped with the appropriate monitoring, error recovery and failure avoidance mechanisms. We employ model-based and data driven service-infrastructure codesign techniques for designing infrastructure and system management deployment. Cloud-based infrastructures pose a special challenge; in public clouds, the unknown deployment of the tenant resources to physical ones and the performance interferences between tenants need continous monitoring and dynamic adaptation.

• In a cloud management framework development project, our experts were involved in the design of model driven and declarative configuration and state management. We also performed measurement-based exploration of cloud resource performance stability and heterogeneity.
• Our experts developed appproaches for the multi-aspect ranking of IaaS platforms.
• Based on historical observations, we devised a capacity planning model for the private Virtual Desktop Infrastructure service of a major financial firm.
• In an EU FP6 project, our experts provided tools and analytic approaches for the dependability management of large, heterogeneous infrastructures.
• We provided a solution for the Hungarian branch of a large international financial institution for the comprehensive modeling of its IT operational processes and characteristics of the underlying IT infrastructrure.

References

A significant amount of Quanopt the R&D expertise of the Quanopt staff was acquired through participation have significant R+D experience in a number of EU FP6 and FP7 projects as researchers of another spin-off company (OptXware).