COMPARATIVE ANALYSIS OF MODELS FOR ESTIMATING THE COMPLEXITY OF IT PROJECTS IN THE ROCKET AND SPACE FIELD: FROM COCOMO II TO FUNCTIONAL POINT ANALYSIS

COMPARATIVE ANALYSIS OF MODELS FOR ESTIMATING THE COMPLEXITY OF IT PROJECTS IN THE ROCKET AND SPACE FIELD: FROM COCOMO II TO FUNCTIONAL POINT ANALYSIS

Ааб Анастасия Валерьевна

аспирант, Сибирский государственный университет науки и технологий имени академика М. Ф. Решетнёва, РФ, г. Красноярск

Abstract. The assessment of the complexity of IT projects at enterprises of the rocket and space industry requires taking into account the specifics of the technological cycle, the cooperative structure, regulated processes and the integration nature of digitalization. The article presents two approaches: the classical COCOMO II model and the function-oriented method of Function Point Analysis. A comparison of their applicability is carried out, taking into account the features of digital transformation in the military-industrial complex. Recommendations on the choice of methods for justifying labor costs in the creation and development of a unified information space are proposed. The priority of functionally oriented models in the context of complex digitalization is substantiated.

Аннотация. Оценка трудоёмкости ИТ-проектов на предприятиях ракетно-космической отрасли требует учёта специфики технологического цикла, кооперационной структуры, регламентированных процессов и интеграционного характера цифровизации. В статье представлены два подхода: классическая модель COCOMO II и функционально-ориентированная методика Function Point Analysis. Проведено сравнение их применимости с учётом особенностей цифровой трансформации в оборонно-промышленном комплексе. Предложены рекомендации по выбору методов для обоснования трудозатрат при создании и развитии единого информационного пространства. Обоснована приоритетность функционально-ориентированных моделей в условиях комплексной цифровизации.

Keywords: Function Point Analysis; COCOMO II; digitalization; automated systems.

Ключевые слова: Function Point Analysis; COCOMO II; цифровизация; автоматизированные системы.

At the current stage of the development of the rocket and space industry ( RSI), the issues of creating a unified information space (UIS) are of strategic importance. Information fragmentation, duplication of data, lack of end-to-end digital traceability of products and weak integration between subsystems negatively affect the manageability, reliability and efficiency of production and design processes.

The development and implementation of EIP are complex IT projects that require precise planning and justification of resources. Of particular importance in this case is the assessment of labor intensity, both at the stage of the feasibility study, and in the process of implementation and control.

Classical methods, such as COCOMO II, were developed based on software development from scratch and rely on code volume indicators. In the context of the RSI, where digitalization proceeds through modernization, adaptation and integration, the use of such methods becomes difficult. Function Point Analysis (FPA) offers an alternative way — an assessment through the number and complexity of functions, which corresponds to the EIP architecture and the process approach in the industry.

IT projects in the rocket and space industry have a number of special features:

• a strict regulatory framework regulates the stages, scope of work, and information security requirements;
• сomplex integration links: PLM, ERP, and MES systems work cooperatively and require secure data exchange;
• long product lifecycle: supporting digital data for 30-50 years requires reliable and scalable solutions;
• using standard platforms with deep refinement: most solutions are not built from scratch, but based on systems such as 1C, SAP, Windchill and others.

Under these conditions, the calculation of labor intensity by lines of code (as in COCOMO II) does not reflect reality: a significant part of the resources is spent on integration, formalization of processes, configuration and validation of solutions. Therefore, evaluation methods should take into account the functional load, not just the software implementation.

The problem of estimating the complexity of software and information systems appeared in the early stages of the development of software engineering. In the initial stages, the main units of measurement were man-days or man-months, and the empirical data accumulated by organizations served as the basis for calculations. However, already in the 1970s, it became obvious that successful project management required unified and formalized evaluation methods that did not depend solely on the team's experience.

At this stage, quantitative models began to form, among which the COCOMO (Constructive Cost Model) model, first introduced by Barry Boehm in 1981, became particularly popular. This model was based on an analysis of more than 60 real-world software development projects and made it possible to link labor costs with the volume of the program in lines of code.

With the development of the IT industry and the transition from monolithic development to integration and flexible approaches, as well as against the background of widespread adoption of corporate information systems (ERP, PLM, CRM), COCOMO in its classic form began to lose relevance. There was a need for a model capable of adequately describing not only the software implementation, but also the functional content of projects.

The COCOMO II model includes three levels:

• Application Composition Model — at the concept stage, estimates based on the number of screens and forms;
• Early Design Model — if there is a preliminary architecture;
• Post-Architecture Model — when the project is fully developed, using cost factors.

The evaluation is based on the number of lines of code (SLOC), adjusted for factors such as reliability, reuse, team experience, process maturity, etc.

However:

• in RSI projects, the amount of code is often unknown until the design stage is completed;
• a significant part of the complexity is spent on integration, configuration, testing, rather than coding;
• the requirements for certification, information security, and GOST do not affect the amount of code, but they significantly increase the actual labor intensity.

Thus, in its classical form, COCOMO II is poorly applicable to the tasks of digitalization in the RSI.

The FPA method evaluates the functional complexity of a system from the user's point of view, based on quantity and complexity:

• inputs (EI),
• outputs (EO),
• Interactive queries (EQ),
• logical files (ELF),
• external interfaces (EIF).

Each element is assigned a weight, the difficulty coefficient is determined, and the total number of function points is calculated.

This is especially important for RSI:

• most projects involve the implementation and configuration of ready—made solutions, where there is little code, but a lot of logic and integration;
• the calculation can be carried out according to the terms of reference, without knowledge of SLOC;
• it is possible to include in the calculation work with standards, integration with PDM, ERP, MES, CD version control, approval routes;
• the model scales and adapts to the specifics of the enterprise.

In practice, FPA allows for a more reliable assessment of projects to create an EIP, build a digital twin, automate cost calculations, planning, and other key functions of the enterprise.

An assessment of the complexity of developing a module for production planning and calculating the cost of a product was carried out at one of the industry enterprises. Have been identified:

• 5 external inputs (item data, routes, fares);
• 4 outputs (reports, cost of production by nodes);
• 3 requests (by deadlines, performers);
• 2 logical files (templates and calculation version);
• 3 external interfaces (PLM, ERP, MES).

The calculation yielded 92 functional points. With a standard of 12 person-hours per point, the total was ~1100 person-hours, which was confirmed by the actual labor costs during the implementation.

COCOMO II remains relevant for evaluating the development of unique software components, but is poorly applicable in the context of digitalization with an emphasis on integration, customization and architecture.

Function Point Analysis adequately reflects the specifics of OT projects in  RSI, focused on building EIP, digital twins, interfacing with PLM/ERP/MES.

The implementation of FPA at the enterprise level makes it possible to standardize estimates of labor costs, improve budgeting accuracy, and simplify digital project management.

For industries like the rocket and space industry, it is advisable to adapt the FPA to GOST and FSTEC requirements, including expanding standard functions and taking into account specific risks.