SOFTWARE BILLS OF MATERIALS (SBOMS) AND ARTIFICIAL INTELLIGENCE BILLS OF MATERIALS (AIBOMS) ARE EMERGING AS CRITICAL SUPPORT FOR MANAGING THE SUPPLY CHAIN OF SOFTWARE AND ARTIFICIAL INTELLIGENCE (AI) SYSTEMS BY IMPROVING TRANSPARENCY, TRACEABILITY, COMPLIANCE, AND MORE. WHILE SBOMS OFFER STRUCTURED, MACHINE-READABLE INVENTORIES OF SOFTWARE COMPONENTS, AIBOMS EXTEND THESE PRINCIPLES TO AI AND AI-ENABLED SYSTEMS BY DOCUMENTING MODELS AND DATASETS. HOWEVER, WE STILL KNOW LITTLE ABOUT THE ADOPTION AND PRACTICAL USAGE OF BOTH SBOM AND AIBOM. THIS THESIS INVESTIGATES THE CURRENT STATE OF PRACTICE OF SBOM AND AIBOM TO SUPPORT THE MANAGEMENT OF THE SUPPLY CHAIN OF SOFTWARE AND AI SYSTEMS, AND CONTRIBUTES TO IMPROVING THE TOOLING OF SBOM AND AIBOM. THROUGH A COMBINATION OF QUALITATIVE AND QUANTITATIVE METHODS—INCLUDING SEMI-STRUCTURED INTERVIEWS WITH PRACTITIONERS, MINING OF OPEN-SOURCE REPOSITORIES, AND A MULTIVOCAL LITERATURE REVIEW—THIS RESEARCH IDENTIFIES LOW AWARENESS AND LIMITED ADOPTION OF SBOM AND AIBOM, AND SIGNIFICANT TECHNICAL LIMITATIONS IN EXISTING SBOM AND AIBOM GENERATORS. TO IMPROVE SUPPORT FOR THE GENERATION OF SBOM AND AIBOM, TWO TOOLS WERE DEVELOPED AND EMPIRICALLY EVALUATED. AIRBORNE ENHANCES SBOMS WITH FINE-GRAINED PROVENANCE DATA BY LINKING SOFTWARE ARTIFACTS TO REUSED CONTENT FROM STACK OVERFLOW. ALOHA GENERATES AIBOMS BY EXTRACTING AND STRUCTURING INFORMATION FROM HUGGING FACE, A MAJOR PLATFORM FOR SHARING AI MODELS AND DATASETS. BOTH TOOLS DEMONSTRATED THE FEASIBILITY OF ENRICHING SBOMS AND AIBOMS THROUGH DEEPER SOFTWARE COMPOSITION ANALYSIS. THE FINDINGS OF THIS THESIS OFFER PRACTICAL AND RESEARCH-ORIENTED IMPLICATIONS. FOR PRACTITIONERS—ESPECIALLY SOFTWARE SUPPLIERS, CHOOSERS, AND OPERATORS—ADOPTING SBOMS AND AIBOMS IS ESSENTIAL FOR MANAGING THE SOFTWARE SUPPLY CHAIN. SUPPLIERS SHOULD INTEGRATE AUTOMATED SBOM/AIBOM GENERATION INTO THEIR PIPELINES, ENSURING COMPLIANCE WITH GUIDELINES, STANDARDS, AND REGULATIONS. CHOOSERS SHOULD DEMAND COMPLETE AND ACCURATE SBOMS/AIBOMS DURING PROCUREMENT, WHILE OPERATORS SHOULD INCORPORATE THEM INTO CI/CD PIPELINES FOR CONTINUOUS RISK MONITORING. FOR RESEARCHERS, FUTURE WORK SHOULD FOCUS ON IMPROVING THE SUPPORT FOR THE GENERATION OF SBOM AND AIBOM, ENHANCING THEIR ACCURACY AND COVERAGE, IMPROVING THE QUALITY OF SUPPORTING DATA SOURCES, AND TRACKING SOFTWARE PROVENANCE MORE EFFECTIVELY.
BILL OF MATERIALS: UNDERSTANDING AND ENHANCING SBOM AND AIBOM TO SUPPORT THE MANAGEMENT OF THE SUPPLY CHAIN OF SOFTWARE AND ARTIFICIAL INTELLIGENCE SYSTEMS / Sabato Nocera , 2025 Dec 18. 38. ciclo, Anno Accademico 2024/25.
BILL OF MATERIALS: UNDERSTANDING AND ENHANCING SBOM AND AIBOM TO SUPPORT THE MANAGEMENT OF THE SUPPLY CHAIN OF SOFTWARE AND ARTIFICIAL INTELLIGENCE SYSTEMS
Nocera, Sabato
2025
Abstract
SOFTWARE BILLS OF MATERIALS (SBOMS) AND ARTIFICIAL INTELLIGENCE BILLS OF MATERIALS (AIBOMS) ARE EMERGING AS CRITICAL SUPPORT FOR MANAGING THE SUPPLY CHAIN OF SOFTWARE AND ARTIFICIAL INTELLIGENCE (AI) SYSTEMS BY IMPROVING TRANSPARENCY, TRACEABILITY, COMPLIANCE, AND MORE. WHILE SBOMS OFFER STRUCTURED, MACHINE-READABLE INVENTORIES OF SOFTWARE COMPONENTS, AIBOMS EXTEND THESE PRINCIPLES TO AI AND AI-ENABLED SYSTEMS BY DOCUMENTING MODELS AND DATASETS. HOWEVER, WE STILL KNOW LITTLE ABOUT THE ADOPTION AND PRACTICAL USAGE OF BOTH SBOM AND AIBOM. THIS THESIS INVESTIGATES THE CURRENT STATE OF PRACTICE OF SBOM AND AIBOM TO SUPPORT THE MANAGEMENT OF THE SUPPLY CHAIN OF SOFTWARE AND AI SYSTEMS, AND CONTRIBUTES TO IMPROVING THE TOOLING OF SBOM AND AIBOM. THROUGH A COMBINATION OF QUALITATIVE AND QUANTITATIVE METHODS—INCLUDING SEMI-STRUCTURED INTERVIEWS WITH PRACTITIONERS, MINING OF OPEN-SOURCE REPOSITORIES, AND A MULTIVOCAL LITERATURE REVIEW—THIS RESEARCH IDENTIFIES LOW AWARENESS AND LIMITED ADOPTION OF SBOM AND AIBOM, AND SIGNIFICANT TECHNICAL LIMITATIONS IN EXISTING SBOM AND AIBOM GENERATORS. TO IMPROVE SUPPORT FOR THE GENERATION OF SBOM AND AIBOM, TWO TOOLS WERE DEVELOPED AND EMPIRICALLY EVALUATED. AIRBORNE ENHANCES SBOMS WITH FINE-GRAINED PROVENANCE DATA BY LINKING SOFTWARE ARTIFACTS TO REUSED CONTENT FROM STACK OVERFLOW. ALOHA GENERATES AIBOMS BY EXTRACTING AND STRUCTURING INFORMATION FROM HUGGING FACE, A MAJOR PLATFORM FOR SHARING AI MODELS AND DATASETS. BOTH TOOLS DEMONSTRATED THE FEASIBILITY OF ENRICHING SBOMS AND AIBOMS THROUGH DEEPER SOFTWARE COMPOSITION ANALYSIS. THE FINDINGS OF THIS THESIS OFFER PRACTICAL AND RESEARCH-ORIENTED IMPLICATIONS. FOR PRACTITIONERS—ESPECIALLY SOFTWARE SUPPLIERS, CHOOSERS, AND OPERATORS—ADOPTING SBOMS AND AIBOMS IS ESSENTIAL FOR MANAGING THE SOFTWARE SUPPLY CHAIN. SUPPLIERS SHOULD INTEGRATE AUTOMATED SBOM/AIBOM GENERATION INTO THEIR PIPELINES, ENSURING COMPLIANCE WITH GUIDELINES, STANDARDS, AND REGULATIONS. CHOOSERS SHOULD DEMAND COMPLETE AND ACCURATE SBOMS/AIBOMS DURING PROCUREMENT, WHILE OPERATORS SHOULD INCORPORATE THEM INTO CI/CD PIPELINES FOR CONTINUOUS RISK MONITORING. FOR RESEARCHERS, FUTURE WORK SHOULD FOCUS ON IMPROVING THE SUPPORT FOR THE GENERATION OF SBOM AND AIBOM, ENHANCING THEIR ACCURACY AND COVERAGE, IMPROVING THE QUALITY OF SUPPORTING DATA SOURCES, AND TRACKING SOFTWARE PROVENANCE MORE EFFECTIVELY.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi Elettronica Nocera Sabato.pdf
accesso aperto
Descrizione: Tesi Elettronica
Tipologia:
Tesi di dottorato
Dimensione
1.57 MB
Formato
Adobe PDF
|
1.57 MB | Adobe PDF | Visualizza/Apri |
|
Abstract Nocera Sabato.pdf
accesso aperto
Descrizione: Abstract
Tipologia:
Tesi di dottorato
Dimensione
20.95 kB
Formato
Adobe PDF
|
20.95 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


