P02 MR RFI 36C10B27Q0010 Attachment 1 M-26-10 Price and Utilization Template 02.xlsx
XLSX spreadsheet 83 KB Posted
- Solicitation number
- 36C10B27Q0010
About this file
This is a Price and Utilization Template (Attachment 1 to M-26-10) designed for federal contractors to report technology spending across multiple product and service categories. The template provides comprehensive data dictionaries and field specifications for reporting on Software (SaaS, IaaS, PaaS, On-Premises), Artificial Intelligence services (LLM/Generative AI APIs, ML Platforms/Training, AI Compute Infrastructure), Cloud services (Reserved Instances, Savings Plans, Committed Use Discounts, Credits), and Hardware purchases. The Software segment requires vendors to report licensing details including category type, license type, unit pricing, quantities procured, and utilization metrics such as assigned users, active users over the last 90 days, and activity thresholds. The AI segment requires detailed tracking of model names, pricing metrics (per token, per API call, per compute-hour, per user), committed versus actual spend, and infrastructure utilization including provisioned instances, average utilization percentages, idle instance counts, and total active versus provisioned instance-hours. The Cloud segment mandates reporting of commitment types, service categories, total committed units or spend, utilized amounts, underutilization gaps, utilization percentages, and credit management including total credits issued, consumed, remaining, expiration dates, and credit sources. The Hardware segment requires simpler reporting of unit pricing, quantities procured, and quantities delivered.
Validation rules are extensive and include requirements that vendor names match SAM.gov legal business names, manufacturer part numbers use only alphanumeric characters with hyphens and periods, product descriptions contain minimum 20 characters with no boilerplate text, pricing fields exclude currency symbols with up to two decimal places, and utilization percentages must fall between 0.00%-100.00%. The template includes controlled dropdown lists for fields such as SW Category Type (IaaS, PaaS, SaaS, OnPrem), License Type (Named User, Concurrent User, Enterprise Agreement, Subscription, Perpetual, Consumption/Metered, Open Source), Unit Measures (Each, Named User License, DBU, Token, Credit, API Call, TB, GB, Hour, Month, Year), AI Service Type (LLM/Generative AI API, ML Platform/Training, AI Compute Infrastructure), Accelerator Type (GPU, TPU, FPGA, CPU-Optimized), and Commitment Type (Reserved Instance, Savings Plan, Committed Use Discount). This template supports federal agencies' cost optimization and FinOps initiatives by enabling standardized tracking of technology spending, utilization efficiency, and identification of underutilized commitments requiring remediation plans when utilization falls below 70% for two consecutive quarters.
View the file
Other files
| File | Type | Posted |
|---|---|---|
| P02 MR RFI 36C10B27Q0010 Attachment A PWS.docx | DOCX document | |
| P02 MR RFI 36C10B27Q0010 Attachment 2 Attestation AI Template_for Awareness.pdf | ||
| 36C10B27Q0010_2.docx | DOCX document | |
| P02 MR RFI 36C10B27Q0010.docx | DOCX document | |
| P02 MR RFI 36C10B27Q0010 Attachment 3 MOU-ISA Template OEM_for Awareness.pdf | ||
| P02 MR RFI 36C10B27Q0010 Attachment A PWS.pdf | ||
| 36C10B27Q0010_1.docx | DOCX document |
On GovTribe
Work with this file
- Download the original file
- Contacts named in this file
- Similar government files
- Ask GovTribe AI about this file
Text version
Software Report
| TDR | TDR | TDR | TDR | TDR | TDR | TDR | TDR | TDR | TDR | |||||
| Contract | Contract | Contract | Product | Product | Product | Product | Product | Quantity and Pricing | Quantity and Pricing | Quantity and Pricing | Quantity and Pricing | Utilization (if available) | Utilization (if available) | Utilization (if available) |
| Ref PIID (if applicable) | PIID | Vendor | Manufacturer Name | Manufacturer Part Number | Description of Deliverable (required) | SW Category Type | SW License Type | Price Paid per Unit (US$) (required) | Quantity Procured | Unit Measure (required) | Total Price (US$) (required) | Quantity Assigned | Quantity Active | Activity Threshold (Days) |
Artificial Intelligence Report
| Contract | Contract | Contract | Product | Product | Product | LLM / Generative AI API Services | LLM / Generative AI API Services | LLM / Generative AI API Services | LLM / Generative AI API Services | LLM / Generative AI API Services | LLM / Generative AI API Services | ML Platform / Training | ML Platform / Training | ML Platform / Training | ML Platform / Training | ML Platform / Training | ML Platform / Training | ML Platform / Training | AI Compute Infrastructure | AI Compute Infrastructure | AI Compute Infrastructure | AI Compute Infrastructure | AI Compute Infrastructure | AI Compute Infrastructure | AI Compute Infrastructure |
| Ref PIID (if applicable) | PIID | Vendor | Manufacturer Name | AI Service Name | AI Service Type | Model Name | Pricing Metric | Unit Price - Input | Unit Price - Output | Committed Spend | Actual Spend | ML Service / Product Name | Pricing Metric | Unit Price | Committed Spend | Actual Spend | Provisioned Training / Inference Instances (count) | Average Utilization of Provisioned Instances (%) | Instance Type | Accelerator Type | Unit Price Per Hour | Provisioned Instances (count) | Total Instance-Hours Provisioned | Total Instance-Hours Actively Used | Number of Idle Instances |
Cloud Report
| Contract | Contract | Contract | Product | Product | Product | Reserved Instance / Commitment Utilization | Reserved Instance / Commitment Utilization | Reserved Instance / Commitment Utilization | Reserved Instance / Commitment Utilization | Reserved Instance / Commitment Utilization | Reserved Instance / Commitment Utilization | Credit / Prepaid Balance Utilization | Credit / Prepaid Balance Utilization | Credit / Prepaid Balance Utilization | Credit / Prepaid Balance Utilization | Credit / Prepaid Balance Utilization |
| Ref PIID (if applicable) | PIID | Vendor | Manufacturer Name | Manufacturer Part Number | Description of Deliverable (required) | Commitment Type | Service Category | Total Committed Units or Spend | Utilized Units or Spend | Underutilized Units or Spend | Utilization Percentage | Total Credits Issued | Credits Consumed | Credits Remaining | Credit Expiration Date | Credit Source |
Hardware Report
| Contract | Contract | Contract | Product | Product | Product | Quantity and Pricing | Quantity and Pricing | Quantity and Pricing | Utilization (if available) |
| Ref PIID (if applicable) | PIID | Vendor | Manufacturer Name | Manufacturer Part Number | Description of Deliverable(required) | Price Paid per Unit (US$) (required) | Quantity Procured | Total Price (US$) (required) | Quantity Delivered |
Data Dictionary Software
| Segment | Field Name | Data Type | Max Length / Format | Description -Validation Rules | Notes |
| Contract | Ref PIID (if applicable) | Text | 50 characters | (if Applicable) PIID of the parent contract vehicle | |
| Contract | PIID | Text | 51 characters | Procurement Identifier | |
| Contract | Vendor | Text | 200 characters | Must match legal business name in SAM.gov for reported UEI; reject if empty | |
| Product | Manufacturer Name | Text | 200 characters | Must not be 'Various' or 'Multiple' for product orders | |
| Product | Manufacturer Part Number | Text | 100 characters | Required for COTS Software; alphanumeric; no special characters except hyphen and period; conditional on Product Category. | |
| Product | Description of Deliverable (required) | Text | 2000 characters | Minimum 20 characters; reject boilerplate/placeholder text (e.g. TBD N/A various); plain language description of product or service delivered; | Current FPDS limitation for the description field is 256 characters |
| Product | SW Category Type | Controlled DropDown | Required | The expected values are IaaS (Infrastructure as a Service), PaaS (Platform as a Service), and SaaS (Software as a Service ) or OnPrem (on-premises software) | |
| Product | SW License Type | Controlled DropDown | 40 characters | Allowed values: Named User | Concurrent User | Named User Plus | Site License | Enterprise Agreement | Subscription | Perpetual | Processor-Based | Core-Based | Capacity / Unit-Based (DBU/Token/Credit) | Consumption / Metered | Open Source | Not Applicable; dropdown | user-based SaaS | processor/core-based (Oracle) | consumption/unit-based (Databricks, AWS, etc); License Type can be used to validate unit of measure field |
Named User Plus is an oracle designation that consist of any possible user or system that can access a database and is not impacted by activation or active user
| Quantity and Pricing | Price Paid per Unit (US$) (required) | Currency | 15 digits 2 decimal | Greater than or equal to $0.00; format: no currency symbol in data field; up to 2 decimal places. | |
| Quantity and Pricing | Quantity Procured | Decimal | 15 digits | Greater than zero; but may use up to 4 decimal places for fractional units (e.g. DBU fractions required for databricks); reject zero or negative values; warn if quantity appears unreasonably large (>1,000,000) but do not reject. For software this is usually number of licenses purchased. | for SaaS it represents Number of licenses purchased |
For IaaS and some PaaS it represents a usage commitment Quantity and Pricing Unit Measure (required) Controlled DropDown 40 characters Allowed values: Each | Named User License | Concurrent User License | CPU License | Core License | Processor License | Instance | DBU (Databricks Unit) | Token | Credit | API Call | TB | GB | Hour | Month | Year | LOT | Seat | Node | Device | Request | NUP; dropdown; must align with License Type Unit Measure options cover processor-based and consumption-based licensing; must be consistent with License Type field
Named User Plus (NUP) is an oracle designation that consist of any possible user or system that can access a database and is not impacted by activation or active user
| Quantity and Pricing | Total Price (US$) (required) | Currency | 15 digits 2 decimal | Must equal Quantity (required) Price Paid per Unit; Can be a system-calculated field. | |
| Utilization (if available) | Quantity Assigned | Integer | 10 digits | Required when License Type is not equal to 'Not Applicable'; valid values are integers from 0 to Qty of Items Sold; represents accounts created in identity provider or vendor tenant regardless of login activity; distinct from active users; also used to show cloud commitments consumed against the dollar value of credits the agency agreed to consume | ASSIGNED: accounts that exist in the system may never have logged in; the number most vendors report by default and most agencies track; gap between this and Active User Count reveals ghost accounts |
Named User Plus is an oracle designation that consist of any possible user or system that can access a database and is not impacted by activation or active user Utilization (if available) Quantity Active (last 90 days) Integer 10 digits Required when License Type is not equal to 'Not Applicable' and when SAM Maturity is "Automated SAM Tool" or "Partial Tooling / Manual Process"; valid values are integers, 0 to Qty of Licenses Assigned; defined as users who authenticated AND performed at least one product action within the reporting calendar month; may be vendor-reported via API or agency SAM portal ACTIVE/CONSUMED: Monthly Active User Count (MAU) the operational utilization metric; number of licenses deemed "active" based on the license activity threshold defined by the agency and programmed into their SAM tool
Named User Plus is an oracle designation that consist of any possible user or system that can access a database and is not impacted by activation or active user Utilization (if available) Activity Threshold (Days) Integer 11 digits The number of days of inactivity to be used to determine quantity active The number of licenses deemed "active" is calculated based on this value. If Activity Threshold = 14 days the quantity of actuve user is all those licensed who have accessed the system in the past 14 days
Data Dictionary AI
| Segment | Field Name | Data Type | Max Length / Format | Description -Validation Rules | Notes |
| Contract | Ref PIID (if applicable) | Text | 50 characters | (if Applicable) PIID of the parent contract vehicle | |
| Contract | PIID | Text | 51 characters | Procurement Identifier | |
| Contract | Vendor | Text | 200 characters | Must match legal business name in SAM.gov for reported UEI; reject if empty | |
| Product | Manufacturer Name | Text | 200 characters | Must not be 'Various' or 'Multiple' for product orders | |
| Product | AI Service Name | Text | 200 characters | The commercial name of the AI service as offered by the vendor; must match the vendor's published service name in their public pricing catalog or API documentation; minimum 3 characters; reject if empty; examples include Amazon Bedrock, Azure OpenAI Service, Google Vertex AI, Anthropic Claude API, OpenAI API | Used to uniquely identify the AI service for cross-agency benchmarking and price analysis; AI service names change frequently with vendor product releases; agencies should verify against current vendor documentation at time of reporting; do not use generic terms such as AI Service or LLM API |
| Product | AI Service Type | Controlled DropDown | 40 characters | Classifies the functional category of the AI service procured; allowed values: LLM / Generative AI API | ML Platform / Training | AI Compute Infrastructure; select the primary service type; if a single line item spans multiple types, split into separate line items for accurate reporting | LLM / Generative AI API: consumption-based access to foundation models via API including text generation, summarization, and embedding services; ML Platform / Training: managed services for training and fine-tuning custom models; AI Compute Infrastructure: raw GPU or TPU compute provisioned for AI workloads without a managed model layer |
| LLM / Generative AI API Services | Model Name | Text | 200 characters | The specific foundation model or model version accessed through the AI service; required when AI Service Type is LLM / Generative AI API; examples include GPT-4o, Claude 3.5 Sonnet, Gemini 1.5 Pro, Llama 3; include the minimum version designation when available | Model name is essential for price reasonableness determination as pricing varies significantly across model tiers within the same service; agencies should record the model name as it appears in vendor API documentation and billing statements; model versioning affects both capability and cost and must be tracked for accurate benchmarking |
| LLM / Generative AI API Services | Pricing Metric | Controlled DropDown | 40 characters | The unit of measure by which the AI service consumption is billed for LLM and Generative AI API services; allowed values: Per Token | Per API Call | Per Compute-Hour | Per User/Seat | Other; must align with the vendor's published pricing structure for the specified model; reject if empty | Per Token: billing based on number of input and output tokens processed; the dominant pricing model for LLM APIs; tokens are typically defined as approximately 4 characters or 0.75 words; Per API Call: flat rate per request regardless of token count; less common for LLMs; Per Compute-Hour: hourly billing for dedicated model endpoints; Per User/Seat: fixed per-user access fee; Other: novel or bespoke pricing structures |
| LLM / Generative AI API Services | Unit Price - Input | Currency | 15 digits 6 decimal | The price charged per unit of input tokens or equivalent input metric submitted to the model; expressed in dollars per 1,000 tokens or per 1 million tokens consistent with vendor pricing conventions; required when Pricing Metric is Per Token; must be greater than or equal to zero. | Input tokens represent the prompt, context window, and system instructions submitted with each API request; input pricing is typically lower than output pricing; agencies should record the exact per-unit price from the vendor invoice or contract; when vendor prices are published per million tokens, divide by 1,000,000 to normalize to per-token unit price for cross-vendor comparison |
| LLM / Generative AI API Services | Unit Price - Output | Currency | 15 digits 6 decimal | The price charged per unit of output tokens or equivalent output metric generated by the model in response to the input; expressed in dollars per 1,000 tokens or per 1 million tokens consistent with vendor pricing conventions; required when Pricing Metric is Per Token; must be greater than or equal to zero. | Output tokens represent the generated response text; output pricing is consistently higher than input pricing across major LLM providers reflecting the greater compute cost of inference generation; agencies should separately track input and output token counts from vendor billing statements to validate this field; the ratio of input to output tokens varies significantly by use case and affects total cost modeling |
| LLM / Generative AI API Services | LLM Committed Spend | Currency | 15 digits 2 decimal | The minimum dollar amount the agency committed to spend on the Large Language Model (LLM) service over the contract term or reporting period in exchange for discounted pricing; required when the line item reflects a committed use or prepaid pricing arrangement; must be greater than zero if present; report as zero or leave blank if the line item is purely on-demand with no commitment | Committed spend arrangements for AI services are analogous to cloud savings plans and may qualify for treatment under proposed FAR 39.106 cloud consumption commitment authority; agencies should document the discount rate achieved relative to on-demand pricing as part of the price reasonableness determination. |
| LLM / Generative AI API Services | LLM Actual Spend | Currency | 15 digits 2 decimal | The total dollar amount billed by the vendor for Large Language Model (LLM) service consumption during the reporting period; must be greater than or equal to zero; cross-validate against vendor invoices. | Actual spend should be reconciled against vendor billing statements monthly; for token-based services actual spend equals the sum of (input tokens consumed x unit price input) plus (output tokens consumed x unit price output); variance between Committed Spend and Actual Spend below 100% of commitment indicates underutilization and should trigger FinOps review |
| ML Platform / Training | ML Service / Product Name | Text | 200 characters | The commercial name of the managed machine learning platform or training service procured; required when AI Service Type is ML Platform / Training; examples include Amazon SageMaker, Azure Machine Learning, Google Vertex AI (training), Databricks MLflow, Domino Data Lab; must match vendor's published product name; reject if empty when AI Service Type is ML Platform / Training | ML platform services are distinct from LLM API services; they provide infrastructure and tooling for building, training, fine-tuning, and deploying custom models; pricing models vary significantly across platforms and may combine compute costs, storage, and platform fees; accurate product name is essential for price benchmarking across agencies |
| ML Platform / Training | Pricing Metric | Controlled DropDown | 40 characters | The unit of measure by which the AI service consumption is billed for ML Platform/Training; allowed values: Per Compute-Hour | Per Inference Request | Per Training Job-Hour | Per Endpoint-Hour | Other | |
| ML Platform / Training | Unit Price | Currency | 15 digits 6 decimal | The per-unit price for the ML platform service as billed by the vendor; the applicable unit depends on the platform's pricing model and must align with the Pricing Metric field; required when AI Service Type is ML Platform / Training; must be greater than or equal to zero | ML platform unit prices may be expressed per compute-hour, per training job-hour, per inference request, or per endpoint-hour depending on the service and usage pattern; agencies should record the unit price as it appears on vendor invoices; Databricks ML services are typically priced in DBUs and should be reported using the DBU unit of measure with the corresponding DBU list or contracted price |
| ML Platform / Training | ML Committed Spend | Currency | 15 digits 2 decimal | The minimum dollar amount the agency committed to spend on the Machine Learning (ML) service over the contract term or reporting period in exchange for discounted pricing; required when the line item reflects a committed use or prepaid pricing arrangement; must be greater than zero if present; report as zero or leave blank if the line item is purely on-demand with no commitment | Committed spend arrangements for AI services are analogous to cloud savings plans and may qualify for treatment under proposed FAR 39.106 cloud consumption commitment authority; agencies should document the discount rate achieved relative to on-demand pricing as part of the price reasonableness determination. |
| ML Platform / Training | ML Actual Spend | Currency | 15 digits 2 decimal | The total dollar amount billed by the vendor for Machine Learning (ML) service consumption during the reporting period; must be greater than or equal to zero; cross-validate against vendor invoices. | Actual spend should be reconciled against vendor billing statements monthly; for token-based services actual spend equals the sum of (input tokens consumed x unit price input) plus (output tokens consumed x unit price output); variance between Committed Spend and Actual Spend below 100% of commitment indicates underutilization. |
| ML Platform / Training | Provisioned Training / Inference Instances (count) | Integer | 10 digits | The number of compute instances provisioned for ML training or inference workloads during the reporting period; required when AI Service Type is ML Platform / Training; must be a positive integer. | Provisioned instances represent reserved compute capacity allocated to ML workloads regardless of actual utilization; distinct from actual usage which is captured in Average Utilization of Provisioned Instances; agencies should monitor provisioned instance counts against active workload requirements and terminate idle instances to reduce cost |
| ML Platform / Training | Average Utilization of Provisioned Instances (%) | Percentage | 5 digits 2 decimal (0.00%–100.00%) | The average percentage of provisioned ML compute capacity actively used during the reporting period; required when AI Service Type is ML Platform / Training and Provisioned Training or Inference Instances count is greater than zero; calculated as average active compute utilization across all provisioned instances during the period. | Low utilization of provisioned ML instances is a primary driver of AI compute waste in federal agencies; utilization data should be sourced from the cloud provider's monitoring service (AWS CloudWatch, Azure Monitor, GCP Cloud Monitoring) or the ML platform's built-in utilization reporting; persistent low utilization should trigger instance rightsizing or consolidation |
| AI Compute Infrastructure | Instance Type | Text | 100 characters | The vendor-specific compute instance type or SKU provisioned for AI compute infrastructure; required when AI Service Type is AI Compute Infrastructure; must match the vendor's published instance type nomenclature; examples include p4d.24xlarge (AWS), Standard_ND96asr_v4 (Azure), a2-megagpu-16g (GCP). | Instance type determines the underlying hardware configuration including CPU, memory, GPU type, and network bandwidth; instance type selection directly affects both performance and cost; agencies should document the business justification for GPU-accelerated instance types when lower-cost alternatives are available; instance type is required for accurate cross-agency cost benchmarking |
| AI Compute Infrastructure | Accelerator Type | Controlled DropDown | 40 characters | The type of hardware accelerator provisioned in the AI compute instance; required when AI Service Type is AI Compute Infrastructure; allowed values: GPU | TPU | FPGA | CPU-Optimized | Other; select the primary accelerator type; if the instance includes multiple accelerator types select the dominant type by compute contribution | GPU: general-purpose graphics processing units used for parallel AI inference and training workloads; the most common accelerator type for LLM fine-tuning and inference; TPU: tensor processing units proprietary to Google Cloud optimized for TensorFlow workloads; FPGA: field-programmable gate arrays used for low-latency inference at the edge; CPU-Optimized: high-core-count CPU instances without discrete GPU used for lightweight inference; accelerator type is a key cost driver and should align with workload requirements |
| AI Compute Infrastructure | Unit Price Per Hour | Currency | 15 digits 4 decimal | The hourly rate charged by the vendor for the provisioned AI compute instance type; required when AI Service Type is AI Compute Infrastructure; must be greater than zero; expressed in US dollars per instance-hour; cross-validate against vendor published on-demand pricing for the specified instance type. | Hourly instance pricing is the standard billing model for IaaS AI compute; agencies operating GPU instances at on-demand rates should evaluate reserved instance or savings plan options to reduce cost; unit price per hour multiplied by total instance-hours provisioned should equal or approximate total billed compute cost for the period; discrepancies may reflect spot instance pricing, sustained use discounts, or billing anomalies requiring reconciliation |
| AI Compute Infrastructure | Provisioned Instances (count) | Integer | 10 digits | The number of AI compute infrastructure instances provisioned during the reporting period; required when AI Service Type is AI Compute Infrastructure; must be a positive integer; represents the maximum number of instances running concurrently at any point during the period. | Provisioned instance count for AI Compute Infrastructure is distinct from Provisioned Training or Inference Instances which applies to ML Platform services; this field captures raw IaaS GPU or TPU instance provisioning; agencies should align provisioned instance counts with actual workload demand and implement auto-scaling where the workload profile supports it to reduce idle compute costs |
| AI Compute Infrastructure | Total Instance-Hours Provisioned | Decimal | 15 digits 2 decimal | The total number of compute instance-hours provisioned during the reporting period, calculated as Provisioned Instances count multiplied by the number of hours in the reporting period; required when AI Service Type is AI Compute Infrastructure; must be greater than zero; system-calculated field. | Total Instance-Hours Provisioned represents maximum possible compute consumption for the period and is the denominator for utilization analysis; for a 90-day quarter: 1 instance provisioned continuously equals 2,160 instance-hours; agencies should report provisioned hours based on actual provisioning records not calendar capacity to account for instances started or stopped during the period |
| AI Compute Infrastructure | Total Instance-Hours Actively Used | Decimal | 15 digits 2 decimal | The total number of compute instance-hours during which the provisioned instances were actively processing AI workloads during the reporting period; required when AI Service Type is AI Compute Infrastructure; must be greater than or equal to zero and less than or equal to Total Instance-Hours Provisioned. | Active use is defined as GPU or accelerator utilization above a minimum threshold (recommended: 10% average GPU utilization) during the instance-hour; idle instances running without active workloads should not be counted as actively used; source data should come from cloud monitoring services (AWS CloudWatch GPU metrics, Azure Monitor, GCP Cloud Monitoring) or vendor utilization reports; the ratio of Instance-Hours Actively Used to Provisioned is the primary AI compute efficiency metric |
| AI Compute Infrastructure | Number of Idle Instances | Integer | 10 digits | The count of instances that were provisioned but had zero or near-zero active usage during the reporing period |
Data Dictionary Cloud
| Segment | Field Name | Data Type | Max Length / Format | Description -Validation Rules | Notes |
| Contract | Ref PIID (if applicable) | Text | 50 characters | (if Applicable) PIID of the parent contract vehicle | |
| Contract | PIID | Text | 51 characters | Procurement Identifier | |
| Contract | Vendor | Text | 200 characters | Must match legal business name in SAM.gov for reported UEI; reject if empty | |
| Product | Manufacturer Name | Text | 200 characters | Must not be 'Various' or 'Multiple' for product orders | |
| Product | Manufacturer Part Number | Text | 100 characters | Alphanumeric; no special characters except hyphen and period; conditional on Product Category. | |
| Product | Description of Deliverable (required) | Text | 2000 characters | Minimum 20 characters; reject boilerplate/placeholder text (e.g. TBD N/A various); plain language description of product or service delivered; | Current FPDS limitation for the description field is 256 characters |
| Reserved Instance / Commitment Utilization | Commitment Type | Controlled DropDown | 40 characters | Identifies the commercial pricing mechanism under which the cloud commitment was established; allowed values: Reserved Instance | Savings Plan | Committed Use Discount | reject if empty for IaaS line items | Reserved Instance: capacity pre-purchased for a fixed term at a discounted rate; Savings Plan: flexible discount tied to a minimum hourly spend commitment regardless of instance type; Committed Use Discount: GCP equivalent of reserved capacity; Other: used for ELA or negotiated commitment structures not covered by standard categories |
| Reserved Instance / Commitment Utilization | Service Category | Controlled DropDown | 40 characters | Classifies the cloud service type covered by the commitment or credit; allowed values: Compute | Database | Storage | Other; select the primary service category consumed under the line item; if multiple categories apply, select the dominant category by spend and note others in Description of Deliverable | Compute: virtual machines, containers, serverless functions; Database: managed relational and NoSQL database services; Storage: object storage, block storage, file storage; Other: networking, security, analytics, and other cloud-native services not classified above |
| Reserved Instance / Commitment Utilization | Total Committed Units or Spend | Currency or Decimal | 15 digits 2 decimal | The total value or quantity the agency committed to consume over the commitment term; for spend-based commitments (Savings Plans) enter the total dollar amount committed; for unit-based commitments (Reserved Instances) enter the total number of instance-hours or equivalent units committed; must be greater than zero; reject if empty | Spend-based: enter dollar amount (e.g., $120,000.00 for a $10,000/month Savings Plan over 12 months); Unit-based: enter instance count or hours; the unit of measure field identifies whether the value is dollars or units; this field establishes the baseline for utilization percentage calculation |
| Reserved Instance / Commitment Utilization | Utilized Units or Spend | Currency or Decimal | 15 digits 2 decimal | The actual units consumed or dollars spent against the commitment during the reporting period; must be greater than or equal to zero and less than or equal to Total Committed Units or Spend. | For Savings Plans: actual spend applied against the committed hourly rate during the period; for Reserved Instances: actual instance-hours consumed; vendor billing consoles (AWS Cost Explorer, Azure Cost Management, GCP Billing) are the authoritative source; agencies should reconcile this field against vendor-issued utilization reports quarterly |
| Reserved Instance / Commitment Utilization | Underutilized Units or Spend | Currency or Decimal | 15 digits 2 decimal | The difference between Total Committed Units or Spend and Utilized Units or Spend; system-calculated as Total Committed minus Utilized; must equal zero or a positive value. | Persistent underutilization triggers remediation requirements; agencies should cross-reference this field against their cloud cost optimization reports; underutilization may result from workload migration, decommissioning, or inaccurate baseline projections at commitment inception |
| Reserved Instance / Commitment Utilization | Utilization Percentage | Percentage | 5 digits 2 decimal (0.00%–100.00%) | Calculated as Utilized Units or Spend divided by Total Committed Units or Spend expressed as a percentage; system-calculated field; must be between 0.00% and 100.00%. | Key FinOps health metric; values below 70% for two consecutive quarters require a remediation plan submitted to agency CIO and CFO; values below 60% for four consecutive quarters require reporting to OMB under applicable FinOps guidance; vendor utilization dashboards are the recommended source for validation |
| Credit / Prepaid Balance Utilization | Total Credits Issued | Currency | 15 digits 2 decimal | The total dollar value of cloud credits issued to the agency under the current contract or ordering instrument; includes negotiated contract credits, promotional credits, and migration incentive credits; must be greater than or equal to zero; required when Credit Source is not null; reject if negative | Credits represent pre-funded cloud consumption value typically issued at contract award or milestone; distinct from commitment pricing discounts; common in large ELA negotiations and migration incentive programs; total credits issued establishes the denominator for credit utilization tracking |
| Credit / Prepaid Balance Utilization | Credits Consumed | Currency | 15 digits 2 decimal | The dollar value of credits applied to actual cloud consumption as of the reporting date; must be greater than or equal to zero and less than or equal to Total Credits Issued. | Credits are typically applied automatically by the vendor billing system before charging the agency; agencies should verify credits consumed against vendor invoices and credit balance statements; credits consumed does not equal total cloud spend — agencies may incur charges beyond the credit balance |
| Credit / Prepaid Balance Utilization | Credits Remaining | Currency | 15 digits 2 decimal | The dollar value of credits not yet consumed as of the reporting date; system-calculated as Total Credits Issued minus Credits Consumed; must be greater than or equal to zero | Agencies should monitor credit burn rate against Credit Expiration Date to avoid forfeiture of unused credits; accelerate consumption planning if credits are at risk of expiring unused; credits remaining represents a financial asset that should be tracked in agency financial systems consistent with applicable accounting standards |
| Credit / Prepaid Balance Utilization | Credit Expiration Date | Date | MM/DD/YYYY | The date on which unused credits expire and become unavailable for consumption; must be a valid calendar date in MM/DD/YYYY format; must be equal to or later than the contract start date; required when Total Credits Issued is greater than zero. | Credit expiration terms vary by vendor and contract; AWS credits typically expire 12 months from issuance; Azure and GCP terms may differ; agencies should negotiate expiration alignment with contract period of performance to minimize forfeiture risk; expired unused credits may represent an Antideficiency Act exposure if originally obligated against appropriated funds |
| Credit / Prepaid Balance Utilization | Credit Source | Controlled DropDown | 40 chars | Identifies the origin or basis of the cloud credits issued; allowed values: Negotiated Contract Credit | Promotional Credit | Migration Incentive | Other; required when Total Credits Issued is greater than zero; reject if empty when credits are present | Negotiated Contract Credit: credits secured as part of ELA or BPA negotiation; Promotional Credit: vendor-issued credits for evaluation, pilot programs, or marketing purposes; Migration Incentive: credits provided to offset workload migration costs from on-premises or competing cloud platforms; Other: credits not covered by standard categories; source classification supports price analysis and future negotiation benchmarking |
Data Dictionary Hardware
| Segment | Field Name | Data Type | Max Length / Format | Description -Validation Rules | Notes |
| Contract | Ref PIID (if applicable) | Text | 50 characters | (if Applicable) PIID of the parent contract vehicle | |
| Contract | PIID | Text | 51 characters | Procurement Identifier | |
| Contract | Vendor | Text | 200 characters | Must match legal business name in SAM.gov for reported UEI; reject if empty | |
| Product | Manufacturer Name | Text | 200 characters | Must not be 'Various' or 'Multiple' for product orders | |
| Product | Manufacturer Part Number | Text | 100 characters | Required for COTS Hardware; alphanumeric; no special characters except hyphen and period; conditional on Product Category. | |
| Product | Description of Deliverable (required) | Text | 2000 characters | Minimum 20 characters; reject boilerplate/placeholder text (e.g. TBD N/A various); plain language description of product or service delivered; | Current FPDS limitation for the description field is 256 characters |
| Quantity and Pricing | Price Paid per Unit (US$) (required) | Currency | 15 digits 2 decimal | Greater than or equal to $0.00; format: no currency symbol in data field; up to 2 decimal places. | |
| Quantity and Pricing | Quantity Procured | Decimal | 15 digits | Greater than zero; but may use up to 4 decimal places for fractional units (e.g. DBU fractions required for databricks); reject zero or negative values; warn if quantity appears unreasonably large (>1,000,000) but do not reject. For physical products it represents the number of hardware units delivered. | for SaaS it represents Number of licenses purchased |
For IaaS and some PaaS it represents a usage commitment
| Quantity and Pricing | Total Price (US$) (required) | Currency | 15 digits 2 decimal | Must equal Quantity (required) Price Paid per Unit; Can be a system-calculated field. | |
| Utilization (if available) | Quantity Delivered | Integer | 10 digits | Represents number of physical units delivered and accepted by the Government; must be an integer greater than or equal to zero. | Used to confirm physical receipt for asset tracking and property management purposes; distinct from Quantity Procured which reflects the order quantity |
Reference_Lists
| CategoryType | License Type | OrderType | Unit of Measure | SAM Maturity | Commitment Type | Service Category | Credit Source | Pricing Metric | AI Service Type | Pricing Metric | Accelerator Type |
| IaaS (Infrastructure as a Service) | User | New Order | Each | Automated SAM Tool | Reserved Instance | Compute | Negotiated Contract Credit | Per 1000 Tokens | LLM / Generative AI API | Per Compute-Hour | GPU |
| PaaS (Platform as a Service) | Site License | Modification | Named User License | Partial Tooling / Manual Processes | Savings Plan | Database | Promotional Credit | Per 1M Tokens | ML Platform / Training | Per Inference Request | TUP |
| SaaS (Software as a Service ) | Enterprise Agreement | Delivery Order | Concurrent User License | No Utilization Data Available | Commited Use Discount | Storage | Migration Incentive | Per API Call | AI Compute Infrastructure | Per Training Job-Hour | FPGA |
| Named User Plus (Oracle) | Per Compute-Hour | ||||||||||
| OnPrem (on-premises HW/SW) | Subscription | Task Order | CPU License | Other | Other | Per Endpoint-Hour | CPU-Optimized | ||||
| Perpetual | BPA Call | Core License | Per User/Seat | Other | Other | ||||||
| Processor-Based | Renewal | Processor License | Other | ||||||||
| Core-Based | Instance | ||||||||||
| Capacity / Unit-Based (DBU/Token/Credit) | DBU (Databricks Unit) | ||||||||||
| Consumption / Metered | Token | ||||||||||
| Open Source | Credit | ||||||||||
| Not Applicable (for hardware) | API Call | ||||||||||
| TB | |||||||||||
| GB | |||||||||||
| Hour | |||||||||||
| Month | |||||||||||
| Year | |||||||||||
| LOT | |||||||||||
| Seat | |||||||||||
| Node | |||||||||||
| Device | |||||||||||
| Request |
File details come from the government source that posted it. Updated .